Lab: Tomotopy#
Package tomotopy [Lee, 2022]
What is tomotopy?#
tomotopy is a Python extension of tomoto (Topic Modeling Tool) which is a Gibbs-sampling based topic model library written in C++.
The current version of tomoto supports several major topic models including
Latent Dirichlet Allocation (
tomotopy.LDAModel)Labeled LDA (
tomotopy.LLDAModel)Partially Labeled LDA (
tomotopy.PLDAModel)Supervised LDA (
tomotopy.SLDAModel)Dirichlet Multinomial Regression (
tomotopy.DMRModel)Generalized Dirichlet Multinomial Regression (
tomotopy.GDMRModel)Hierarchical Dirichlet Process (
tomotopy.HDPModel)Hierarchical LDA (
tomotopy.HLDAModel)Multi Grain LDA (
tomotopy.MGLDAModel)Pachinko Allocation (
tomotopy.PAModel)Hierarchical PA (
tomotopy.HPAModel)Correlated Topic Model (
tomotopy.CTModel)Dynamic Topic Model (
tomotopy.DTModel)Pseudo-document based Topic Model (
tomotopy.PTModel).
Getting Started#
You can install tomotopy easily using pip. (https://pypi.org/project/tomotopy/)
pip install --upgrade pip
pip install tomotopy
After installing, you can start tomotopy by just importing.
import tomotopy as tp
print(tp.isa) # prints 'avx2', 'avx', 'sse2' or 'none'
Here is a sample code for simple LDA training of texts from ‘sample.txt’ file.
import tomotopy as tp
mdl = tp.LDAModel(k=20)
for line in open('sample.txt'):
mdl.add_doc(line.strip().split())
for i in range(0, 100, 10):
mdl.train(10)
print('Iteration: {}\tLog-likelihood: {}'.format(i, mdl.ll_per_word))
for k in range(mdl.k):
print('Top 10 words of topic #{}'.format(k))
print(mdl.get_topic_words(k, top_n=10))
mdl.summary()
Model Save and Load#
tomotopy provides save and load method for each topic model class,
so you can save the model into the file whenever you want, and re-load it from the file.
import tomotopy as tp
mdl = tp.HDPModel()
for line in open('sample.txt'):
mdl.add_doc(line.strip().split())
for i in range(0, 100, 10):
mdl.train(10)
print('Iteration: {}\tLog-likelihood: {}'.format(i, mdl.ll_per_word))
# save into file
mdl.save('sample_hdp_model.bin')
# load from file
mdl = tp.HDPModel.load('sample_hdp_model.bin')
for k in range(mdl.k):
if not mdl.is_live_topic(k): continue
print('Top 10 words of topic #{}'.format(k))
print(mdl.get_topic_words(k, top_n=10))
# the saved model is HDP model,
# so when you load it by LDA model, it will raise an exception
mdl = tp.LDAModel.load('sample_hdp_model.bin')
When you load the model from a file, a model type in the file should match the class of methods.
Documents in the Model and out of the Model#
We can use Topic Model for two major purposes. The basic one is to discover topics from a set of documents as a result of trained model, and the more advanced one is to infer topic distributions for unseen documents by using trained model.
We named the document in the former purpose (used for model training) as document in the model, and the document in the later purpose (unseen document during training) as document out of the model.
In tomotopy, these two different kinds of document are generated differently.
A document in the model can be created by tomotopy.LDAModel.add_doc method.
add_doc can be called before tomotopy.LDAModel.train starts.
In other words, after train called, add_doc cannot add a document into the model because the set of document used for training has become fixed.
To acquire the instance of the created document, you should use tomotopy.LDAModel.docs like:
mdl = tp.LDAModel(k=20)
idx = mdl.add_doc(words)
if idx < 0: raise RuntimeError("Failed to add doc")
doc_inst = mdl.docs[idx]
# doc_inst is an instance of the added document
A document out of the model is generated by tomotopy.LDAModel.make_doc method. make_doc can be called only after train starts.
If you use make_doc before the set of document used for training has become fixed, you may get wrong results.
Since make_doc returns the instance directly, you can use its return value for other manipulations.
mdl = tp.LDAModel(k=20)
# add_doc ...
mdl.train(100)
doc_inst = mdl.make_doc(unseen_doc) # doc_inst is an instance of the unseen document
Inference for Unseen Documents#
If a new document is created by tomotopy.LDAModel.make_doc, its topic distribution can be inferred by the model.
Inference for unseen document should be performed using tomotopy.LDAModel.infer method.
mdl = tp.LDAModel(k=20)
# add_doc ...
mdl.train(100)
doc_inst = mdl.make_doc(unseen_doc)
topic_dist, ll = mdl.infer(doc_inst)
print("Topic Distribution for Unseen Docs: ", topic_dist)
print("Log-likelihood of inference: ", ll)
The infer method can infer only one instance of tomotopy.Document or a list of instances of tomotopy.Document.
See more at tomotopy.LDAModel.infer.
Corpus and transform#
Every topic model in tomotopy has its own internal document type.
A document can be created and added into suitable for each model through each model’s add_doc method.
However, trying to add the same list of documents to different models becomes quite inconvenient,
because add_doc should be called for the same list of documents to each different model.
Thus, tomotopy provides tomotopy.utils.Corpus class that holds a list of documents.
tomotopy.utils.Corpus can be inserted into any model by passing as argument corpus to __init__ or add_corpus method of each model.
So, inserting tomotopy.utils.Corpus just has the same effect to inserting documents the corpus holds.
Some topic models requires different data for its documents.
For example, tomotopy.DMRModel requires argument metadata in str type,
but tomotopy.PLDAModel requires argument labels in List[str] type.
Since tomotopy.utils.Corpus holds an independent set of documents rather than being tied to a specific topic model,
data types required by a topic model may be inconsistent when a corpus is added into that topic model.
In this case, miscellaneous data can be transformed to be fitted target topic model using argument transform.
See more details in the following code:
from tomotopy import DMRModel
from tomotopy.utils import Corpus
corpus = Corpus()
corpus.add_doc("a b c d e".split(), a_data=1)
corpus.add_doc("e f g h i".split(), a_data=2)
corpus.add_doc("i j k l m".split(), a_data=3)
model = DMRModel(k=10)
model.add_corpus(corpus)
# You lose `a_data` field in `corpus`,
# and `metadata` that `DMRModel` requires is filled with the default value, empty str.
assert model.docs[0].metadata == ''
assert model.docs[1].metadata == ''
assert model.docs[2].metadata == ''
def transform_a_data_to_metadata(misc: dict):
return {'metadata': str(misc['a_data'])}
# this function transforms `a_data` to `metadata`
model = DMRModel(k=10)
model.add_corpus(corpus, transform=transform_a_data_to_metadata)
# Now docs in `model` has non-default `metadata`, that generated from `a_data` field.
assert model.docs[0].metadata == '1'
assert model.docs[1].metadata == '2'
assert model.docs[2].metadata == '3'
Pining Topics using Word Priors#
Since version 0.6.0, a new method tomotopy.LDAModel.set_word_prior has been added. It allows you to control word prior for each topic.
For example, we can set the weight of the word ‘church’ to 1.0 in topic 0, and the weight to 0.1 in the rest of the topics by following codes.
This means that the probability that the word ‘church’ is assigned to topic 0 is 10 times higher than the probability of being assigned to another topic.
Therefore, most of ‘church’ is assigned to topic 0, so topic 0 contains many words related to ‘church’.
This allows to manipulate some topics to be placed at a specific topic number.
import tomotopy as tp
mdl = tp.LDAModel(k=20)
# add documents into `mdl`
# setting word prior
mdl.set_word_prior('church', [1.0 if k == 0 else 0.1 for k in range(20)])
Examples#
Install or upgrade of ekorpkit#
Note
Install ekorpkit package first.
Set logging level to Warning, if you don’t want to see verbose logging.
If you run this notebook in Colab, set Hardware accelerator to GPU.
!pip install -U –pre ekorpkit[topic]
exit()
from ekorpkit import eKonf
eKonf.setLogger("WARNING")
print("version:", eKonf.__version__)
print("is notebook?", eKonf.is_notebook())
print("is colab?", eKonf.is_colab())
print("environment variables:")
eKonf.print(eKonf.env().dict())
data_dir = "../data/topic_models"
version: 0.1.39.post0.dev7
is notebook? True
is colab? False
environment variables:
{'CUDA_DEVICE_ORDER': None,
'CUDA_VISIBLE_DEVICES': None,
'EKORPKIT_CONFIG_DIR': '/workspace/projects/ekorpkit-book/config',
'EKORPKIT_DATA_DIR': None,
'EKORPKIT_LOG_LEVEL': 'WARNING',
'EKORPKIT_PROJECT': 'ekorpkit-book',
'EKORPKIT_WORKSPACE_ROOT': '/workspace',
'KMP_DUPLICATE_LIB_OK': 'TRUE',
'NUM_WORKERS': 230}
Load a dataset#
cfg = eKonf.compose("path")
cfg.cache.uri = "https://github.com/entelecheia/ekorpkit-book/raw/main/assets/data/us_equities_news_sampled.zip"
data = eKonf.load_data("us_equities_news_sampled.parquet", cfg.cached_path)
data.text[0]
'Investing com Asian stock markets were broadly lower for a second day on Thursday as weak U S data on durable goods orders added to concerns over the global growth outlook while concerns over declining corporate profits also weighed During late Asian trade Hong Kong s Hang Seng Index tumbled 1 55 Australia s ASX 200 Index dipped 0 1 while Japan s Nikkei 225 Index shed 0 7 The Nikkei came further off a one year closing high hit earlier in the week as investors cashed in ahead of the Japanese fiscal year end March is the final month of Japan s fiscal year and market participants have expected many funds to lock in profits from a meteoric 19 rally in the January to March period after shedding more than 13 in April to December Exporters which have gained sharply in the first quarter on the back a weakening yen declined Automakers Toyota and Nissan slumped 1 65 and 1 8 respectively while consumer electronics giant Sony retreated 1 5 On the upside Sharp saw shares jump 6 7 extending the previous day s 15 rally following reports that Taiwan s Hon Hai Precision Industry is buying 10 of the Japanese electronics manufacturer for JPY66 91 billion with the two to form a tie up in liquid crystal display production Elsewhere shares in Hong Kong came under pressure amid lingering fears over a hard landing in China and worries over declining corporate profits PICC Property Casualty China s biggest non life insurer saw shares drop 4 after reporting 2011 net income rose to CNY8 03 billion missing expectations for income of CNY8 8 billion China Shipping Container Lines fell 1 55 after the nation s second largest container carrier reported a loss of CNY2 74 billion last year wider than the average estimate of CNY2 63 billion amid rising fuel costs and declining demand Port operator China Merchants Holdings declined 1 7 after saying annual profit dropped 5 2 due to costs from a harbor deal Hong Kong traded Chinese banks were also broadly weaker with Industrial and Commercial Bank of China and Bank of China down 2 4 and 2 3 respectively ahead of their 2011 earnings reports later on Thursday Raw material producers also contributed to losses amid concerns over the global growth outlook Copper mining giant Jiangxi Copper Company dropped 3 4 Aluminum Corporation of China or CHALCO slumped 2 4 while shares in oil majors PetroChina and CNOOC fell 2 1 and 3 8 respectively But shares in Australia outperformed regional equities for a second day remaining close to a four month high The index was weighed by a 7 decline in Leighton Holdings the country s largest construction company after saying its underlying profit in the year ending December 31 will be between AUD400 million and AUD450 million below market expectations citing increased costs due to wet weather and lower than expected productivity Meanwhile European stock markets were mildly lower after the open supported by hopes that euro zone leaders will increase the size of the European debt firewall to combat the fiscal crisis The EURO STOXX 50 shed 0 3 France s CAC 40 dipped 0 1 Germany s DAX fell 0 25 while London s FTSE 100 edged 0 15 lower Later in the day Germany was to publish official data on employment change while the U S was to release official data on initial jobless claims'
LDA Basics#
LDA class provides Latent Dirichlet Allocation(LDA) topic model and its implementation is based on following papers:
Blei, D.M., Ng, A.Y., &Jordan, M.I. (2003).Latent dirichlet allocation.Journal of machine Learning research, 3(Jan), 993 - 1022.
Newman, D., Asuncion, A., Smyth, P., &Welling, M. (2009).Distributed algorithms for topic models.Journal of Machine Learning Research, 10(Aug), 1801 - 1828.
import tomotopy as tp
save_path = eKonf.join_path(data_dir, "lda_basic.mdl")
mdl = tp.LDAModel(tw=tp.TermWeight.ONE, min_cf=10, rm_top=10, k=20)
for n, line in enumerate(data["text"][:100]):
ch = line.strip().split()
mdl.add_doc(ch)
mdl.burn_in = 100
mdl.train(0)
print(
"Num docs:",
len(mdl.docs),
", Vocab size:",
len(mdl.used_vocabs),
", Num words:",
mdl.num_words,
)
print("Removed top words:", mdl.removed_top_words)
for i in range(0, 100, 10):
mdl.train(10)
print("Iteration: {}\tLog-likelihood: {}".format(i, mdl.ll_per_word))
mdl.save(save_path, full=True)
Show code cell output
Num docs: 100 , Vocab size: 972 , Num words: 33695
Removed top words: ['the', 'to', 'of', 'and', 'in', 'a', 'is', 'for', 's', 'on']
Iteration: 0 Log-likelihood: -7.3065742522854
Iteration: 10 Log-likelihood: -7.055851222156576
Iteration: 20 Log-likelihood: -6.9786549585511946
Iteration: 30 Log-likelihood: -6.918256985219584
Iteration: 40 Log-likelihood: -6.906922808788813
Iteration: 50 Log-likelihood: -6.875851555821591
Iteration: 60 Log-likelihood: -6.860287052386768
Iteration: 70 Log-likelihood: -6.837860183047866
Iteration: 80 Log-likelihood: -6.833746569962225
Iteration: 90 Log-likelihood: -6.823284888087085
mdl.summary()
Show code cell output
<Basic Info>
| LDAModel (current version: 0.12.3)
| 100 docs, 33695 words
| Total Vocabs: 8855, Used Vocabs: 972
| Entropy of words: 6.35945
| Entropy of term-weighted words: 6.35945
| Removed Vocabs: the to of and in a is for s on
|
<Training Info>
| Iterations: 100, Burn-in steps: 100
| Optimization Interval: 10
| Log-likelihood per word: -6.82328
|
<Initial Parameters>
| tw: TermWeight.ONE
| min_cf: 10 (minimum collection frequency of words)
| min_df: 0 (minimum document frequency of words)
| rm_top: 10 (the number of top words to be removed)
| k: 20 (the number of topics between 1 ~ 32767)
| alpha: [0.1] (hyperparameter of Dirichlet distribution for document-topic, given as a single `float` in case of symmetric prior and as a list with length `k` of `float` in case of asymmetric prior.)
| eta: 0.01 (hyperparameter of Dirichlet distribution for topic-word)
| seed: 3080440490 (random seed)
| trained in version 0.12.3
|
<Parameters>
| alpha (Dirichlet prior on the per-document topic distributions)
| [0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
| 0.1 0.1]
| eta (Dirichlet prior on the per-topic word distribution)
| 0.01
|
<Topics>
| #0 (4616) : has have be an it
| #1 (979) : trade market that from growth
| #2 (1177) : he He Trump his said
| #3 (1842) : as that The oil from
| #4 (1204) : It that The by some
| #5 (2673) : quarter 1 year 3 company
| #6 (783) : 1 ratio market margin value
| #7 (2400) : with this that from their
| #8 (1257) : its high as this while
| #9 (2030) : percent 1 0 by 2
| #10 (1354) : I gold t you silver
| #11 (1154) : revenue PepsiCo In billion that
| #12 (2895) : The earnings year P S
| #13 (1123) : Goldman by Reuters up last
| #14 (1125) : companies are NYSE they more
| #15 (1115) : with 7 by sales The
| #16 (2150) : said it not that company
| #17 (1204) : S at U said investors
| #18 (1736) : Zacks Rank growth has expected
| #19 (878) : down demand China it level
|
for k in range(mdl.k):
print("Topic #{}".format(k))
for word, prob in mdl.get_topic_words(k):
print("\t", word, prob, sep="\t")
Show code cell output
Topic #0
has 0.043887220323085785
have 0.041293028742074966
be 0.040644481778144836
an 0.03415900468826294
it 0.03286190703511238
are 0.03156481310725212
will 0.031132448464632034
at 0.02745734713971615
or 0.027241162955760956
than 0.023782242089509964
Topic #1
trade 0.047546323388814926
market 0.04248927906155586
that 0.03844364359974861
from 0.03439800813794136
growth 0.03439800813794136
8 0.030352376401424408
4 0.029340967535972595
quarter 0.02731814980506897
25 0.02428392320871353
results 0.022261105477809906
Topic #2
he 0.05730922520160675
He 0.04045604541897774
Trump 0.037085410207509995
his 0.03624275326728821
said 0.03455743566155434
but 0.031186800450086594
was 0.028658824041485786
by 0.026973506435751915
with 0.026973506435751915
will 0.025288190692663193
Topic #3
as 0.05886959284543991
that 0.05130905658006668
The 0.046988748013973236
oil 0.03240770846605301
from 0.03186766803264618
two 0.028627438470721245
out 0.025927245616912842
last 0.02484716847538948
production 0.022687014192342758
over 0.019986823201179504
Topic #4
It 0.05109086260199547
that 0.04944303259253502
The 0.04861912131309509
by 0.04037998989224434
some 0.037908248603343964
also 0.03296476975083351
global 0.02966911531984806
stock 0.028845202177762985
price 0.024725638329982758
world 0.023077810183167458
Topic #5
quarter 0.04585271701216698
1 0.044734448194503784
year 0.03914311155676842
3 0.03839759901165962
company 0.038024842739105225
2 0.03280625864863396
million 0.03131523355841637
from 0.02907869778573513
billion 0.02460562437772751
4 0.023860113695263863
Topic #6
1 0.0706554651260376
ratio 0.044164396822452545
market 0.044164396822452545
margin 0.04037995636463165
value 0.03659551963210106
company 0.029026644304394722
debt 0.02524220570921898
currency 0.021457767114043236
13 0.02019628696143627
total 0.02019628696143627
Topic #7
with 0.05478229746222496
this 0.05395232513546944
that 0.05270737037062645
from 0.047727540135383606
their 0.02905316837131977
just 0.023243363946676254
see 0.022828377783298492
as 0.02241339161992073
investment 0.018678518012166023
time 0.017433559522032738
Topic #8
its 0.05605816841125488
high 0.05132152512669563
as 0.045795440673828125
this 0.03553271293640137
while 0.0292171910405159
NASDAQ 0.0292171910405159
week 0.02842774987220764
day 0.026059428229928017
new 0.02526998706161976
markets 0.0244805496186018
Topic #9
percent 0.08923283219337463
1 0.05589492991566658
0 0.05491440370678902
by 0.04020649939775467
2 0.0367746576666832
with 0.024027807638049126
yen 0.02255701646208763
after 0.02255701646208763
expected 0.020595964044332504
than 0.018634909763932228
Topic #10
I 0.09166838228702545
gold 0.06160355731844902
t 0.040338192135095596
you 0.03740504011511803
silver 0.029338866472244263
Gold 0.027872290462255478
out 0.02493913657963276
think 0.020539406687021255
do 0.019806118682026863
money 0.019806118682026863
Topic #11
revenue 0.05758258327841759
PepsiCo 0.03695906326174736
In 0.03524043411016464
billion 0.03524043411016464
that 0.03352180868387222
The 0.029225243255496025
EPS 0.026647305116057396
are 0.024928677827119827
growth 0.024928677827119827
investment 0.02149142511188984
Topic #12
The 0.057496074587106705
earnings 0.048889391124248505
year 0.03615150600671768
P 0.02582348883152008
S 0.02444641850888729
3 0.02375788427889347
2 0.02169227972626686
up 0.02134801261126995
average 0.02100374549627304
stock 0.01824960671365261
Topic #13
Goldman 0.03708771616220474
by 0.03532205522060394
Reuters 0.030025072395801544
up 0.029142240062355995
last 0.025610920041799545
security 0.023845259100198746
were 0.022962426766753197
July 0.022079596295952797
further 0.020313935354351997
Inc 0.020313935354351997
Topic #14
companies 0.08020481467247009
are 0.06522314250469208
NYSE 0.05464784428477287
they 0.04142872244119644
more 0.03702234849333763
000 0.03525979816913605
about 0.029972150921821594
like 0.02820960246026516
some 0.0238032303750515
index 0.022921955212950706
Topic #15
with 0.05157728120684624
7 0.05157728120684624
by 0.05068816989660263
sales 0.0471317321062088
The 0.046242620795965195
year 0.03824062645435333
Inc 0.024903977289795876
Amazon 0.024903977289795876
5 0.02401486597955227
at 0.022236647084355354
Topic #16
said 0.051863204687833786
it 0.04815902188420296
not 0.03704646974802017
that 0.03519437834620476
company 0.0342683307826519
was 0.032879263162612915
who 0.030564147979021072
at 0.0277860090136528
its 0.02454484812915325
would 0.024081824347376823
Topic #17
S 0.08322348445653915
at 0.0683930441737175
U 0.06674521416425705
said 0.03131694346666336
investors 0.02966911531984806
dollar 0.027197375893592834
The 0.023077810183167458
seen 0.022253897041082382
rate 0.02060607075691223
up 0.019782157614827156
Topic #18
Zacks 0.10655201971530914
Rank 0.06072566285729408
growth 0.049841899424791336
has 0.034375499933958054
expected 0.025783058255910873
company 0.025783058255910873
over 0.022918909788131714
currently 0.019481932744383812
Estimate 0.019481932744383812
you 0.01890910230576992
Topic #19
down 0.04619699716567993
demand 0.041691072285175323
China 0.037185147404670715
it 0.030426261946558952
level 0.028173299506306648
after 0.028173299506306648
inflation 0.027046818286180496
consumer 0.025920337066054344
Google 0.020287929102778435
from 0.020287929102778435
LDA Visualization#
This example shows how to perform a Latent Dirichlet Allocation using tomotopy and visualize the result.
import tomotopy as tp
import nltk
import re
import numpy as np
import pyLDAvis
from nltk.corpus import stopwords
porter_stemmer = nltk.PorterStemmer().stem
english_stops = set(porter_stemmer(w) for w in stopwords.words("english"))
pat = re.compile("^[a-z]{2,}$")
corpus = tp.utils.Corpus(
tokenizer=tp.utils.SimpleTokenizer(porter_stemmer),
stopwords=lambda x: x in english_stops or not pat.match(x),
)
corpus.process(d.lower() for d in data["text"][:100])
# save preprocessed corpus for reuse
save_path = eKonf.join_path(data_dir, "preprocessed_corpus.cps")
corpus.save(save_path)
mdl = tp.LDAModel(min_df=5, rm_top=20, k=30, corpus=corpus)
mdl.train(0)
print(
"Num docs:{}, Num Vocabs:{}, Total Words:{}".format(
len(mdl.docs), len(mdl.used_vocabs), mdl.num_words
)
)
print("Removed Top words: ", *mdl.removed_top_words)
Num docs:100, Num Vocabs:1074, Total Words:22386
Removed Top words: year compani stock market earn said zack price percent share quarter expect growth report billion revenu trade investor nyse million
# Let's train the model
for i in range(0, 100, 10):
print("Iteration: {:04}, LL per word: {:.4}".format(i, mdl.ll_per_word))
mdl.train(10)
print("Iteration: {:04}, LL per word: {:.4}".format(1000, mdl.ll_per_word))
mdl.summary()
Show code cell output
Iteration: 0000, LL per word: -12.02
Iteration: 0010, LL per word: -7.564
Iteration: 0020, LL per word: -7.357
Iteration: 0030, LL per word: -7.263
Iteration: 0040, LL per word: -7.193
Iteration: 0050, LL per word: -7.167
Iteration: 0060, LL per word: -7.159
Iteration: 0070, LL per word: -7.144
Iteration: 0080, LL per word: -7.138
Iteration: 0090, LL per word: -7.128
Iteration: 1000, LL per word: -7.122
<Basic Info>
| LDAModel (current version: 0.12.3)
| 100 docs, 22386 words
| Total Vocabs: 5139, Used Vocabs: 1074
| Entropy of words: 6.62188
| Entropy of term-weighted words: 6.62188
| Removed Vocabs: year compani stock market earn said zack price percent share quarter expect growth report billion revenu trade investor nyse million
|
<Training Info>
| Iterations: 100, Burn-in steps: 0
| Optimization Interval: 10
| Log-likelihood per word: -7.12162
|
<Initial Parameters>
| tw: TermWeight.ONE
| min_cf: 0 (minimum collection frequency of words)
| min_df: 5 (minimum document frequency of words)
| rm_top: 20 (the number of top words to be removed)
| k: 30 (the number of topics between 1 ~ 32767)
| alpha: [0.1] (hyperparameter of Dirichlet distribution for document-topic, given as a single `float` in case of symmetric prior and as a list with length `k` of `float` in case of asymmetric prior.)
| eta: 0.01 (hyperparameter of Dirichlet distribution for topic-word)
| seed: 3743484366 (random seed)
| trained in version 0.12.3
|
<Parameters>
| alpha (Dirichlet prior on the per-document topic distributions)
| [0.32200417 0.60077 0.21077627 0.06970649 0.4447674 0.5434947
| 0.5469532 0.45518324 0.6338666 0.5719722 0.8117618 0.47851452
| 0.5830207 0.5280389 0.46882564 0.58426094 0.5307975 0.526721
| 0.66027516 0.20483024 0.817219 0.5032047 0.20158361 0.4396458
| 0.43283358 0.6222955 0.43815643 0.29211614 0.32323366 0.4832791 ]
| eta (Dirichlet prior on the per-topic word distribution)
| 0.01
|
<Topics>
| #0 (486) : currenc high forecast chang way
| #1 (1118) : make think would come anoth
| #2 (525) : yen trump japan presid reuter
| #3 (403) : gold silver metal mine see
| #4 (692) : fund yield larg etf countri
| #5 (753) : high manag move new last
| #6 (877) : week rose fall ahead juli
| #7 (765) : rank consensu posit cloud per
| #8 (788) : invest may time import major
| #9 (926) : buy estim current see sell
| #10 (1130) : industri past hold rank term
| #11 (686) : level valu move follow risk
| #12 (817) : increas rate new demand govern
| #13 (890) : dollar bank futur sinc like
| #14 (733) : secur firm accord activist group
| #15 (780) : estim per ep best beat
| #16 (888) : sector goldman invest us season
| #17 (702) : day averag drop show corpor
| #18 (969) : strong increas result continu cent
| #19 (279) : use system organ center network
| #20 (1220) : also busi one last grow
| #21 (566) : month data china gain rise
| #22 (483) : oil product mexico energi unit
| #23 (703) : index higher close nasdaq fell
| #24 (529) : plan car would state media
| #25 (1070) : inc product nasdaq technolog top
| #26 (739) : back could call pay comment
| #27 (623) : oper ratio total margin dividend
| #28 (571) : sale retail store amazon consum
| #29 (675) : peopl financi look equiti seen
|
topic_term_dists = np.stack([mdl.get_topic_word_dist(k) for k in range(mdl.k)])
doc_topic_dists = np.stack([doc.get_topic_dist() for doc in mdl.docs])
doc_topic_dists /= doc_topic_dists.sum(axis=1, keepdims=True)
doc_lengths = np.array([len(doc.words) for doc in mdl.docs])
vocab = list(mdl.used_vocabs)
term_frequency = mdl.used_vocab_freq
prepared_data = pyLDAvis.prepare(
topic_term_dists,
doc_topic_dists,
doc_lengths,
vocab,
term_frequency,
start_index=0, # tomotopy starts topic ids with 0, pyLDAvis with 1
sort_topics=False, # IMPORTANT: otherwise the topic_ids between pyLDAvis and tomotopy are not matching!
)
pyLDAvis.display(prepared_data)
save_dir = "../../../assets/extra"
filename = "lda_basic.html"
save_path = eKonf.join_path(save_dir, filename)
pyLDAvis.save_html(prepared_data, save_path)
from IPython.display import display, HTML
display(HTML(f"<a href={save_path} target='_blank'>{filename}</a>"))
LDA coherence#
This example shows how to perform a Latent Dirichlet Allocation and calculate coherence of the results.
# calculate coherence using preset
for preset in ("u_mass", "c_uci", "c_npmi", "c_v"):
coh = tp.coherence.Coherence(mdl, coherence=preset)
average_coherence = coh.get_score()
coherence_per_topic = [coh.get_score(topic_id=k) for k in range(mdl.k)]
print("==== Coherence : {} ====".format(preset))
print("Average:", average_coherence, "\nPer Topic:", coherence_per_topic)
print()
Show code cell output
==== Coherence : u_mass ====
Average: -1.2919665389931534
Per Topic: [-1.17386418557316, -0.8680202687611864, -1.9011274512865044, -1.047667736559903, -0.8366067935289703, -1.0824157815953828, -1.0848464209397464, -1.042256352533198, -2.040907684969454, -0.9648597265277666, -0.6370809895484615, -1.0602086759498055, -0.9300058656023976, -0.8755223027370869, -3.8756595693460896, -0.9561534903362163, -1.0182683509407169, -1.0452619935102012, -0.8683999688876635, -1.6542543949045063, -0.9976812325252717, -1.3517324017228247, -1.4340397599847348, -1.1336003900525302, -3.3037643601664413, -0.9452589069885053, -1.3000326329631926, -1.1188280272579738, -1.1003419572424118, -1.1103284968522937]
==== Coherence : c_uci ====
Average: -4.878004424663513
Per Topic: [-1.4440664765834577, -2.9780619010239433, -8.708752137868833, -3.0050339171614286, -3.9756801762639395, -5.2244161333617765, -6.687473406288814, -4.228681125865242, -7.003535776093569, -3.3012711354385487, -0.9000521475110517, -3.7900076427042646, -3.7450666628395886, -3.05733166371, -7.122763804824756, -4.240481845103053, -5.210568908088681, -3.822197030576038, -4.042555006747717, -8.432800473204122, -4.129114331246837, -4.856754672889186, -4.192296480531112, -5.375999981840818, -10.244442932288125, -1.8611091058190008, -7.7178064083191105, -2.0541820205768704, -5.6412571244580745, -9.346372310677454]
==== Coherence : c_npmi ====
Average: -0.1325481235208962
Per Topic: [0.017405643466831665, -0.06713514820217745, -0.250626658245578, -0.04635542135098983, -0.12268777150029593, -0.1647925543989872, -0.19153348461822214, -0.11479310813952832, -0.20325291559000797, -0.05933672775478992, -0.01312994262040823, -0.10370859429506546, -0.12495710347546096, -0.08055725886583803, -0.18367815044046804, -0.1056187907799229, -0.16950989984821221, -0.1013101627858306, -0.12513122278003186, -0.23018273615527657, -0.13565862607035908, -0.14425460236335147, -0.029011488535021748, -0.16785973124959508, -0.3279976915096982, -0.03667616244336371, -0.25424020821444454, 0.04555992891405496, -0.15709031459299314, -0.32832280118185425]
==== Coherence : c_v ====
Average: 0.5564370689820499
Per Topic: [0.5011142507195473, 0.5537001609802246, 0.6300975859165192, 0.6528063505887985, 0.5177711397409439, 0.4293242484331131, 0.593023756146431, 0.687737911939621, 0.48875993937253953, 0.6143026351928711, 0.4760036665014923, 0.4910168096423149, 0.4035642951726913, 0.5804461449384689, 0.6122414320707321, 0.5381382368505001, 0.5088414877653122, 0.5503150254487992, 0.49591032415628433, 0.6861237555742263, 0.4793075716122985, 0.5599398672580719, 0.6356056213378907, 0.5512014642357826, 0.6266018450260162, 0.5769483402371407, 0.5394160777330399, 0.6937022149562836, 0.5420901507139206, 0.4770597591996193]
import itertools
# calculate coherence using custom combination
for seg, cm, im in itertools.product(
tp.coherence.Segmentation, tp.coherence.ConfirmMeasure, tp.coherence.IndirectMeasure
):
coh = tp.coherence.Coherence(
mdl, coherence=(tp.coherence.ProbEstimation.DOCUMENT, seg, cm, im)
)
average_coherence = coh.get_score()
coherence_per_topic = [coh.get_score(topic_id=k) for k in range(mdl.k)]
print("==== Coherence : {}, {}, {} =S===".format(repr(seg), repr(cm), repr(im)))
print("Average:", average_coherence, "\nPer Topic:", coherence_per_topic)
print()
Show code cell output
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.NONE: 0> =S===
Average: 0.1351874610643829
Per Topic: [0.11101613159001103, 0.14175311097344043, 0.16218523657111322, 0.18403090427713625, 0.13375353018989322, 0.11100999661663936, 0.11403467499327492, 0.19470537423940926, 0.09631062974767758, 0.16263582197968346, 0.13661738585142094, 0.11390028382623481, 0.11258613327854435, 0.1902236282755808, 0.1258461538435896, 0.1319174081261988, 0.11790130382631658, 0.101183821270789, 0.13940305217600987, 0.16222103513433575, 0.11832280151689759, 0.09788931234235808, 0.1759731325287265, 0.1419625983771921, 0.12283456255415204, 0.11983944773489327, 0.1397846799755621, 0.17328171317695196, 0.12248752611724914, 0.10001244082020401]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.351438074447317
Per Topic: [0.33813833362526363, 0.430223838157124, 0.2877488086620967, 0.459748269058764, 0.37483677003118726, 0.3659698486328125, 0.3136667526430554, 0.4313059644566642, 0.3343541645341449, 0.451196633776029, 0.3850690140078465, 0.40851979123221505, 0.29808357678767705, 0.5845632533232371, 0.1891543977169527, 0.32126699541178017, 0.3908944098485841, 0.3035161207119624, 0.34076554398569797, 0.3225314312924941, 0.3825223781996303, 0.29911605285273657, 0.3335316630287303, 0.3914314246426026, 0.2338264191316234, 0.35029977937018786, 0.3607129294011328, 0.31518845558166503, 0.2355661774882012, 0.30939303582741157]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.DICE: 2> =S===
Average: -0.8293424091943761
Per Topic: [0.007356486366026932, 0.09002556798255278, -19.3119921805544, 0.0981799298690425, 0.041860342708726725, 0.058977754372689456, -0.42121727665782804, -1.2304922645497653, 0.02389199184771213, 0.030123389098379347, 0.04062372690273656, 0.0743200851811303, -0.056963336705747575, 0.18184349636236827, -1.326329428785377, -0.430772896633587, 0.013559697485632367, -0.07364538814872504, -0.016037708872722253, -0.8799245460269352, 0.06188205188243753, -0.0832902544281549, -0.006324527770306708, 0.02566142504931324, -0.44272010181254395, -0.1809079928504717, -0.06870913652399192, -0.8275042325607501, -0.256899498630729, -0.014847449427987967]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.05052750938504081
Per Topic: [0.028326774704166585, 0.1131498378701508, -0.3143009778112173, 0.11495007860163847, 0.05273476496545805, 0.07745150006893609, -0.19890510874489944, -0.4506096941108505, 0.03240897646173835, 0.05276763825159934, 0.07799616557442479, 0.10300765294167731, 0.029785138120253882, 0.22646935797399945, -0.30842970315780904, -0.21479199138056074, 0.04015165230052339, -0.06290167171715035, 0.03329444641454352, -0.3777039244822744, 0.0837298682696807, -0.048015411863000027, 0.0031752510759462086, 0.06049309560718636, -0.2972231205811517, -0.10548304048522066, -0.04337195058696024, -0.14943984682775205, -0.09097979786909288, 0.016438758864791857]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.NONE: 0> =S===
Average: 1.7045623505630207
Per Topic: [1.5266369490971852, 1.5463879113428056, 2.5005275414844204, 2.6211654329045535, 1.615207291260249, 1.4244162026935911, 1.524343961603905, 1.8431163771644157, 1.4003549336881114, 1.482846216572118, 1.3435245777160014, 1.4048138054696506, 1.3353415497716667, 1.7875376955323294, 1.9259544157948747, 1.4914559665787048, 1.5159353629304715, 1.3652218150352202, 1.4493477040020373, 2.7839088670144436, 1.3712513309499346, 1.4650723061795412, 2.5894590816212744, 1.6079433553160218, 1.9996923130945299, 1.4098756572162499, 1.7141295931091425, 2.025746019068962, 1.6136944791990455, 1.451961803479181]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.8551825504170523
Per Topic: [0.8568254854944017, 0.906697412331899, 0.7271665891011556, 0.8156261616282993, 0.8707838190926446, 0.8924511830012004, 0.8620413064956665, 0.8509526835547553, 0.8582547346750895, 0.9164952185418871, 0.937135538789961, 0.9030599276224772, 0.9020455718040467, 0.9209719644652472, 0.7183065354824066, 0.8972178141276042, 0.8740266468789842, 0.8944442338413663, 0.9104682432280646, 0.6893860750728183, 0.926681931813558, 0.8576761537128025, 0.7724587930573358, 0.8919851157400344, 0.6979990694257948, 0.9129396239916484, 0.8691527287165324, 0.8243044892946879, 0.8304114023844401, 0.8675100591447619]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.DICE: 2> =S===
Average: 0.3953467755295613
Per Topic: [0.3952705932988061, 0.4116775261031257, 0.3457771937052409, 0.3738812698258294, 0.3972695622179243, 0.4121724492973752, 0.3927939838833279, 0.39261747068829006, 0.3986912210782369, 0.4234406484497918, 0.4311494575606452, 0.4138374083571964, 0.41399666402075025, 0.4178920825322469, 0.35133889416853586, 0.4098752948972914, 0.4021126025252872, 0.41012402772903445, 0.41826743086179097, 0.3370403753386603, 0.4250685731569926, 0.3981595748000675, 0.35959198474884035, 0.4010718471474118, 0.3426904241243998, 0.41851362850930957, 0.3994131624698639, 0.3785300427012973, 0.3875945382648044, 0.4005433334244622]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.6596688153346381
Per Topic: [0.6563803288671706, 0.7020963006549411, 0.5323186298211415, 0.6005045566293928, 0.6619628310203552, 0.7054342985153198, 0.6484619776407877, 0.6504347576035394, 0.6657542082998488, 0.736786593331231, 0.759814174969991, 0.7082306557231479, 0.710816662841373, 0.7194733553462558, 0.5525827613141802, 0.6960519154866537, 0.675995961825053, 0.6966831141048008, 0.7229824145634969, 0.5109996477762858, 0.7403956360287136, 0.6643118752373589, 0.5646324700779385, 0.6727087252669864, 0.5272161457273695, 0.720808560318417, 0.6659841272566054, 0.612063941028383, 0.6380366259151035, 0.6701412068472968]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.NONE: 0> =S===
Average: 833248193.1610017
Per Topic: [2.0082930761066575, 2.0649914529938123, 5.536606056036479, 4074074078.1978583, 2.210945653270672, 1.931628999164661, 2.125470572564088, 2.950515135257185, 1.8025983892937552, 2.373465908815218, 1.8405752408315796, 1.8027322625961777, 1367816093.6748478, 2.8848589893698096, 19555555557.625587, 2.0319740312759507, 1.9730925319655515, 1.6820690483526008, 1.9618860918242997, 4.003495981109249, 1.859194891827816, 1.9271989339005193, 3.968215875053943, 2.4390567261732707, 2.937843376827051, 1.7903929410806272, 2.227756137638969, 2.9174473414104023, 2.1420887517385103, 1.937361064931399]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.7402534365717239
Per Topic: [0.7901608188947041, 0.843464023537106, 0.4551342276028461, 0.3149481895897124, 0.7948817253112793, 0.8571844696998596, 0.8122045384513007, 0.7541576802730561, 0.8069455133544075, 0.8641168276468912, 0.8905021548271179, 0.8615469124582079, 0.7176932009588028, 0.8644522918595208, 0.27564658059061264, 0.831760479344262, 0.8138297478357951, 0.8520880222320557, 0.8684459182951185, 0.42272013458940716, 0.8856943421893649, 0.8053802145851983, 0.63527436653773, 0.8174244178666009, 0.4379139787207047, 0.8728516340255738, 0.8224291814698114, 0.6472298730578687, 0.7715571337276035, 0.8199644976192051]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.DICE: 2> =S===
Average: 0.33317976239575325
Per Topic: [0.3563797414302826, 0.3753472553359138, 0.2330093636364418, 0.1232157862403148, 0.35636063549253677, 0.3826446420616574, 0.3597119006845686, 0.3548059927092658, 0.36521188616752626, 0.394514090485043, 0.40643646253479854, 0.38578488296932645, 0.3188276247833031, 0.3827680640750461, 0.10762037602072624, 0.3783985793590546, 0.3688957737551795, 0.3822804285420312, 0.39233422146903146, 0.18592197238293587, 0.39835565818680657, 0.36646339164839853, 0.29996280868848163, 0.35952309105131364, 0.1976712567250935, 0.3922528200679355, 0.3683128794034322, 0.2780546369890838, 0.3555277672078874, 0.3687988817691803]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.5316045413682643
Per Topic: [0.5580196784602272, 0.6070075160927243, 0.3320080794477474, 0.19004649233998394, 0.5593037558926477, 0.625640649927987, 0.5640980223814647, 0.5553434246116215, 0.5791655408011542, 0.6549664517243703, 0.6875664088461134, 0.6306196133295695, 0.5326852523605549, 0.6239841527409024, 0.16353459524120556, 0.6113076812691158, 0.5893126898341708, 0.6207689881324768, 0.6512066079510583, 0.2668755257019132, 0.6637576699256897, 0.5820300175084008, 0.4329821487267812, 0.5680274128913879, 0.29369887570285214, 0.6470509966214498, 0.5842362410492368, 0.4285337212423949, 0.557279497385025, 0.5870785329076978]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.NONE: 0> =S===
Average: 0.5405859769669961
Per Topic: [0.5826810441083398, 0.6660489351224613, 0.359891652097845, 1.321096407409169, 0.6701883879830325, 0.5319959935158121, 0.5658087696202627, 0.9456190446093845, -0.11162484937707866, 0.7352332243367302, 0.5597643057419147, 0.522965590300311, 0.7758243685876923, 0.8836533972493942, -1.1621559621911957, 0.6014894707978905, 0.5812047616908983, 0.4615187886617018, 0.6076949115341027, 1.0887713264147822, 0.5399529510350565, 0.4875531804415069, 1.1033524638471561, 0.6779175245241125, -1.0429912454491495, 0.5440375487491753, 0.6974779167686911, 0.904638576448713, 0.6177343167893081, 0.5002365076418617]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.11935449063389972
Per Topic: [0.09104066548170522, 0.06104483656365321, 0.3093050652080112, 0.28175778368280996, 0.08212855363470023, 0.008025267011382514, 0.12080522253006873, 0.1865225060661841, 0.030552612939532587, 0.1300905320340664, 0.059253980701517626, 0.041128822967099646, -0.013953080879420869, 0.09569137090713614, 0.17172210053023365, 0.13141210740238118, 0.09200627624264193, 0.09485439706542012, 0.03239099182925808, 0.3284517904122671, 0.029139023562897312, 0.09320585189982214, 0.20915542462219794, 0.14141440639117112, 0.19086722037431578, 0.07778486348171201, 0.1352714304941603, 0.24153346777929982, 0.08869124026872062, 0.039339987812046374]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.DICE: 2> =S===
Average: 9.764055218641406
Per Topic: [1.4785268584887186, 8.029365785916646, 0.9702214002609253, 1.8041651871469286, 6.375461806191338, -0.7050536228550806, 1.4947402609719171, 2.4129498998324075, 3.9956230812602573, 5.715696952078077, -3.1508254329363505, -3.0207321961720783, -0.4067401882674959, 1.2127181026670668, 1.1320184853341844, 211.9972156604131, 1.2562444779607984, 1.7112665004200405, -0.5146351165241665, 0.9359890752368503, -4.61125920481152, 5.525151567988925, 39.07899330457052, 1.171814935737186, 0.9255641990237766, 2.0271761390897964, 1.5980050881703696, 1.2008336848682828, 1.1204772697554695, 2.160682597425249]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.JACCARD: 3> =S===
Average: -20.825088196727965
Per Topic: [-2.7075881520907084, -2.272716283135944, 3.2274680429034763, 8.821884269184537, -4.040207725101047, -2.5036019414663313, -0.882104041841295, 18.635549034012687, -1.0613703462812636, -3.0882180836465625, -1.289473260111279, -2.3885018050670626, -2.2382670289940303, -2.2146502597464455, -0.25286320050557454, -6.070304331514571, -3.884779371155633, -2.4781847609413994, -2.379622534248564, -604.0762025992076, -1.6511220229996575, -0.15254030227661133, -5.448412187894186, -3.880208917458852, 4.612057495117187, -3.180569440788693, -3.9969701210657758, -0.6741435077455309, 6.395788433154424, -3.6327709509266746]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.NONE: 0> =S===
Average: 0.2978416590213351
Per Topic: [0.3780543494180079, 0.41296956919561634, 0.18980654629698537, 0.7691373178155088, 0.43306236075249366, 0.309001147774482, 0.3558471752030721, 0.5755005634088165, -0.2125246165440904, 0.3698390002038444, 0.28614394288966083, 0.3204577164651875, 0.2715209622206862, 0.5262686189508873, -1.423538163168902, 0.3624002845095718, 0.3701383246947222, 0.2877505226455023, 0.3516867899079114, 0.8617703727076176, 0.29600534443070325, 0.3136588599661505, 0.839832037893, 0.4166274363483794, -0.9223808370094703, 0.3299575706066555, 0.48694471889678254, 0.6346445910731934, 0.4248759551503945, 0.31979130793668337]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.17994302767120876
Per Topic: [0.12805647044959995, 0.11670115184194098, 0.32550658281478617, 0.3205361928480367, 0.1428330622962676, 0.17555796836192408, 0.16637061934840555, 0.2242692696241041, 0.06496669764682236, 0.2734027441797985, 0.23776455167163577, 0.1406386292203226, 0.10335524227573639, 0.2027769971690658, 0.1911078766376401, 0.19030399062256848, 0.15196374676500757, 0.13059370080526506, 0.12559629707700676, 0.3441504937079218, 0.1384310316397912, 0.1333530227472592, 0.21875141391323671, 0.17642224971204995, 0.20602956772264508, 0.14699623237570955, 0.16957532687520144, 0.2709827545409401, 0.10145181030060889, 0.07984513494496544]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.DICE: 2> =S===
Average: 2.7841405712333676
Per Topic: [1.666050275001261, 0.6319933338297739, 0.9500268008973863, 1.0386896597014532, 2.250735392794013, 0.2756188680003915, 0.17857569058736164, 0.9083827204174466, -2.425627151255806, 0.8707619605792893, -0.2592729784548283, -0.19907247605216172, 0.802290514152911, 0.8496030672556825, 1.077572684817844, 1.2845817714929582, 3.02141981518103, 1.3063109305169847, -0.03761112979716725, 0.9118721087773641, -0.33338644554217656, 1.1887123200628493, 1.0431044512324863, 1.445079586075412, 0.9786576999558343, 61.388298433356816, 1.0510625998179117, 1.1705422348446317, 1.6549388404418197, -1.1656944416877297]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.JACCARD: 3> =S===
Average: 7.0628127242821375
Per Topic: [-9.166540891263221, -2.342065739466084, -6.15212869114346, -43.87733449406094, 56.44401636720738, -2.6669036716222765, 1.8036339667108323, 0.8181499573919508, -3.2101934199945794, -3.841389800608158, -1.0703579567372798, -2.519921213326355, -4.007283565733168, -2.817425860464573, -3.0735806226730347, -1.983945485121674, -5.066192412128051, -8.380187131961186, -2.8874885131087566, -4.33076155450609, -1.4770257814062966, -2.8912580755021837, 289.9129321826829, -3.833444407582283, -8.179676280419033, -4.677306281195746, -2.820743430985345, 2.5057655387454565, -4.367011500419014, -3.959949502845605]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.NONE: 0> =S===
Average: 0.1766937747568545
Per Topic: [0.14824682586212107, 0.18708438146379341, 0.20281956968249248, 0.24425282842966542, 0.17419189085101458, 0.14265440707643529, 0.14944404510495848, 0.2521999987572534, 0.10834079577106827, 0.2163301201716118, 0.19242175789761948, 0.15242108698649745, 0.15412284717444524, 0.24714972156952805, 0.11141293530537302, 0.17405880215656602, 0.15648258888424169, 0.13498831971804393, 0.1852731795046233, 0.2408658489817811, 0.15769056713778093, 0.12939673533270415, 0.25091810534198355, 0.18260028518696703, 0.12661959749425444, 0.15883896199371866, 0.18870800407669353, 0.2339521853385075, 0.16486529111142129, 0.1324615583424696]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.30799426429340176
Per Topic: [0.2577278188119332, 0.36455040506811603, 0.27655373497141733, 0.37316074293080925, 0.30523637334505715, 0.31315051799433097, 0.25819794668091667, 0.3904047669635879, 0.226201621008416, 0.4634987852639622, 0.43473703985412915, 0.3618507874508699, 0.286240578111675, 0.5096372005012301, 0.16657824232760404, 0.31665071715186865, 0.29665989984447755, 0.26487721796664926, 0.34197703417804504, 0.3128788981690175, 0.3643739732810193, 0.22974988263514307, 0.2334528881435593, 0.34796334778269133, 0.1884039598180809, 0.3390489612188604, 0.2937325588117043, 0.29894404822132653, 0.18890943971152108, 0.2344785405840311]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.DICE: 2> =S===
Average: -6.696977050953278
Per Topic: [-0.501081607511474, -0.058775322056478924, -5.950174129340383, -1.0146991021827692, -0.19650192408718997, -0.5972039008306133, 68.61222103497842, -0.5676436835382548, -0.43249382376670836, -0.2505337163909442, 0.004097158461809159, -0.45448160244462393, 2.5815661487686965, 0.10029054445525011, 3.3434767007827757, -12.810671669674209, -0.8501870264609654, -0.34551202019469607, 13.165061561514934, 0.3700122098128001, -0.013202836861213048, 0.8229714737584194, -0.018339197296235295, -0.1298573174048215, -0.5295765637523598, -270.0585608153294, -0.7114471904726491, 5.680398739212089, 0.04282868843939569, -0.14129233918696021]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.JACCARD: 3> =S===
Average: 2.1071468107418276
Per Topic: [-0.2321656532999542, -0.009382950059241718, -1.7773478090763093, -0.3373492727159626, -0.12773477082244225, -0.14153379789657064, -0.58121748691968, -0.5889425377361477, -0.20039150988062224, -0.03989723877360423, 0.0805715151131153, -0.027872106507937942, -0.122680844925344, 0.12404205043696695, -1.1835544238487878, -0.5044889359683212, -0.24931146192053955, -0.21346006609706414, -0.1409227941185236, 9.27308953884575, 0.0325390866647164, -0.3664802716837989, -0.6178307306435373, -0.08279434137253297, 63.20406588969959, -0.19907079886438117, -0.30779310458650194, -0.6767237674030993, -0.5674934085458517, -0.2034636748385512]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.NONE: 0> =S===
Average: -1.2272521961226583
Per Topic: [-1.1493610416512177, -0.9092812064780683, -1.970326895536801, -1.209571909913137, -1.0353222293274311, -0.9238462153413554, -1.1178646906898124, -0.9071476677119651, -1.7523576293806506, -0.7866509199609911, -0.6362178372900401, -0.9782684757568221, -0.8257068116299219, -0.7863431660657688, -3.6965884747288964, -0.9244933878967375, -1.0690041290183774, -0.9893611453220097, -0.795524288275124, -1.504613733026627, -0.7923858074358401, -1.1512319143641574, -1.2850092927096146, -0.9458022725617946, -3.164209136644416, -0.8555574890591362, -1.1133245636868838, -1.171646059376813, -1.200905406979172, -1.1696420858601588]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.4489587392133696
Per Topic: [0.4172645108567344, 0.3680855777528551, 0.47551360925038655, 0.46617349055078294, 0.42326092951827576, 0.6414405763149261, 0.38452149331569674, 0.4236492501364814, 0.45179317759142984, 0.4734817637337579, 0.513688063621521, 0.4954831686284807, 0.5364161756303575, 0.4674039426777098, 0.3921919425328573, 0.45504247943560283, 0.40978858404689367, 0.34588676674498453, 0.4799913730886247, 0.5026300244861179, 0.48717684911357034, 0.3698685281806522, 0.45938059820069205, 0.48731302552753025, 0.452699175145891, 0.40655995971626707, 0.4119377543528875, 0.42373811536365086, 0.4009397059679031, 0.44544156491756437]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.DICE: 2> =S===
Average: 0.7170875298093867
Per Topic: [0.7426078796386719, 0.7015290180842082, 0.7337837298711141, 0.7339953780174255, 0.7182757245169745, 0.7243074602550931, 0.7232087903552585, 0.7609699527422588, 0.7437359650929769, 0.704906948407491, 0.6405402686860826, 0.6788614100880093, 0.7135660039054023, 0.6997373965051439, 0.7442686306105719, 0.7021779139836629, 0.7368649337026808, 0.714368814892239, 0.682016827000512, 0.7392514573203193, 0.7062499708599514, 0.7357237325774298, 0.7286302831437853, 0.7034077525138855, 0.7649813916948106, 0.6894622564315795, 0.6908912089135911, 0.7162305646472507, 0.7261502756012811, 0.7119239542219374]
==== Coherence : <Segmentation.ONE_ONE: 1>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.JACCARD: 3> =S===
Average: 2.9189544194716004
Per Topic: [3.515099125438266, 2.5720635890960692, 3.0092657830980087, 3.2807204246521, 2.874766953786214, 3.7289518859651354, 2.8617791864607067, 3.7924955818388195, 3.4049933248096043, 2.7408284081353083, 1.896372185813056, 2.4040002981821695, 3.1123066107432047, 2.618385820918613, 3.2927805900573732, 2.509185202916463, 3.3346821069717407, 2.619035180409749, 2.6232512129677668, 3.124668386247423, 2.6490831957923042, 3.215196074379815, 2.9361532661649914, 2.5999829795625473, 3.845041184955173, 2.3929948170979816, 2.3027957253985933, 2.6045406103134154, 2.966953155729506, 2.740259716245863]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.NONE: 0> =S===
Average: 0.1259989069056579
Per Topic: [0.1061245305061154, 0.14334851156873857, 0.16778857356312876, 0.20327777777305223, 0.1617922072094629, 0.08897302425204513, 0.11153553343618199, 0.1664350771921802, 0.06624612605006143, 0.13548263442620365, 0.1378050723416717, 0.10646782095923937, 0.10722655288680988, 0.16845434930938016, 0.10646723646540263, 0.12095094119498716, 0.1218401308467869, 0.09246942280151692, 0.12911488191024367, 0.14767602767328883, 0.09694586040870846, 0.08211303852897013, 0.15206043846975117, 0.11745118766041761, 0.10556810678833571, 0.10622589196313859, 0.11827787994978789, 0.17570182084653996, 0.12563702444193925, 0.11050952574565061]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.351438074447317
Per Topic: [0.33813833362526363, 0.430223838157124, 0.2877488086620967, 0.459748269058764, 0.37483677003118726, 0.3659698486328125, 0.3136667526430554, 0.4313059644566642, 0.3343541645341449, 0.451196633776029, 0.3850690140078465, 0.40851979123221505, 0.29808357678767705, 0.5845632533232371, 0.1891543977169527, 0.32126699541178017, 0.3908944098485841, 0.3035161207119624, 0.34076554398569797, 0.3225314312924941, 0.3825223781996303, 0.29911605285273657, 0.3335316630287303, 0.3914314246426026, 0.2338264191316234, 0.35029977937018786, 0.3607129294011328, 0.31518845558166503, 0.2355661774882012, 0.30939303582741157]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.DICE: 2> =S===
Average: -0.8293424091943761
Per Topic: [0.007356486366026932, 0.09002556798255278, -19.3119921805544, 0.0981799298690425, 0.041860342708726725, 0.058977754372689456, -0.42121727665782804, -1.2304922645497653, 0.02389199184771213, 0.030123389098379347, 0.04062372690273656, 0.0743200851811303, -0.056963336705747575, 0.18184349636236827, -1.326329428785377, -0.430772896633587, 0.013559697485632367, -0.07364538814872504, -0.016037708872722253, -0.8799245460269352, 0.06188205188243753, -0.0832902544281549, -0.006324527770306708, 0.02566142504931324, -0.44272010181254395, -0.1809079928504717, -0.06870913652399192, -0.8275042325607501, -0.256899498630729, -0.014847449427987967]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.05052750938504081
Per Topic: [0.028326774704166585, 0.1131498378701508, -0.3143009778112173, 0.11495007860163847, 0.05273476496545805, 0.07745150006893609, -0.19890510874489944, -0.4506096941108505, 0.03240897646173835, 0.05276763825159934, 0.07799616557442479, 0.10300765294167731, 0.029785138120253882, 0.22646935797399945, -0.30842970315780904, -0.21479199138056074, 0.04015165230052339, -0.06290167171715035, 0.03329444641454352, -0.3777039244822744, 0.0837298682696807, -0.048015411863000027, 0.0031752510759462086, 0.06049309560718636, -0.2972231205811517, -0.10548304048522066, -0.04337195058696024, -0.14943984682775205, -0.09097979786909288, 0.016438758864791857]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.NONE: 0> =S===
Average: 1.7045623505630207
Per Topic: [1.5266369490971856, 1.5463879113428052, 2.500527541484421, 2.6211654329045517, 1.6152072912602486, 1.4244162026935918, 1.5243439616039047, 1.8431163771644155, 1.4003549336881116, 1.4828462165721186, 1.3435245777160023, 1.4048138054696502, 1.3353415497716672, 1.7875376955323292, 1.9259544157948747, 1.4914559665787055, 1.5159353629304713, 1.3652218150352202, 1.4493477040020368, 2.7839088670144436, 1.3712513309499348, 1.4650723061795419, 2.5894590816212744, 1.607943355316022, 1.9996923130945297, 1.4098756572162503, 1.7141295931091431, 2.025746019068962, 1.6136944791990455, 1.451961803479182]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.8551825504170523
Per Topic: [0.8568254854944017, 0.906697412331899, 0.7271665891011556, 0.8156261616282993, 0.8707838190926446, 0.8924511830012004, 0.8620413064956665, 0.8509526835547553, 0.8582547346750895, 0.9164952185418871, 0.937135538789961, 0.9030599276224772, 0.9020455718040467, 0.9209719644652472, 0.7183065354824066, 0.8972178141276042, 0.8740266468789842, 0.8944442338413663, 0.9104682432280646, 0.6893860750728183, 0.926681931813558, 0.8576761537128025, 0.7724587930573358, 0.8919851157400344, 0.6979990694257948, 0.9129396239916484, 0.8691527287165324, 0.8243044892946879, 0.8304114023844401, 0.8675100591447619]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.DICE: 2> =S===
Average: 0.3953467755295613
Per Topic: [0.3952705932988061, 0.4116775261031257, 0.3457771937052409, 0.3738812698258294, 0.3972695622179243, 0.4121724492973752, 0.3927939838833279, 0.39261747068829006, 0.3986912210782369, 0.4234406484497918, 0.4311494575606452, 0.4138374083571964, 0.41399666402075025, 0.4178920825322469, 0.35133889416853586, 0.4098752948972914, 0.4021126025252872, 0.41012402772903445, 0.41826743086179097, 0.3370403753386603, 0.4250685731569926, 0.3981595748000675, 0.35959198474884035, 0.4010718471474118, 0.3426904241243998, 0.41851362850930957, 0.3994131624698639, 0.3785300427012973, 0.3875945382648044, 0.4005433334244622]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.6596688153346381
Per Topic: [0.6563803288671706, 0.7020963006549411, 0.5323186298211415, 0.6005045566293928, 0.6619628310203552, 0.7054342985153198, 0.6484619776407877, 0.6504347576035394, 0.6657542082998488, 0.736786593331231, 0.759814174969991, 0.7082306557231479, 0.710816662841373, 0.7194733553462558, 0.5525827613141802, 0.6960519154866537, 0.675995961825053, 0.6966831141048008, 0.7229824145634969, 0.5109996477762858, 0.7403956360287136, 0.6643118752373589, 0.5646324700779385, 0.6727087252669864, 0.5272161457273695, 0.720808560318417, 0.6659841272566054, 0.612063941028383, 0.6380366259151035, 0.6701412068472968]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.NONE: 0> =S===
Average: 1014644532.1307074
Per Topic: [2.0716905148356126, 2.045420831647486, 6.179741581835933, 8148148151.5435505, 1.96388427810584, 2.1567399942720877, 2.1689504559035084, 3.295981720050152, 2.031448664028519, 2.761785608553607, 1.817477707696492, 1.8154245281843358, 2735632185.589626, 3.2561513797552215, 19555555557.634724, 2.1183098679418264, 1.9320395422644892, 1.742234440291489, 2.0395844394243032, 4.133499290822854, 2.070774345681312, 2.1017326507876506, 4.527260013181598, 2.805270916317296, 3.061376765367417, 1.8785028352814224, 2.348135626366958, 2.8691604139030074, 2.207386240153904, 1.7533559920550403]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.7402534365717239
Per Topic: [0.7901608188947041, 0.843464023537106, 0.4551342276028461, 0.3149481895897124, 0.7948817253112793, 0.8571844696998596, 0.8122045384513007, 0.7541576802730561, 0.8069455133544075, 0.8641168276468912, 0.8905021548271179, 0.8615469124582079, 0.7176932009588028, 0.8644522918595208, 0.27564658059061264, 0.831760479344262, 0.8138297478357951, 0.8520880222320557, 0.8684459182951185, 0.42272013458940716, 0.8856943421893649, 0.8053802145851983, 0.63527436653773, 0.8174244178666009, 0.4379139787207047, 0.8728516340255738, 0.8224291814698114, 0.6472298730578687, 0.7715571337276035, 0.8199644976192051]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.DICE: 2> =S===
Average: 0.33317976239575325
Per Topic: [0.3563797414302826, 0.3753472553359138, 0.2330093636364418, 0.1232157862403148, 0.35636063549253677, 0.3826446420616574, 0.3597119006845686, 0.3548059927092658, 0.36521188616752626, 0.394514090485043, 0.40643646253479854, 0.38578488296932645, 0.31882762478330307, 0.3827680640750461, 0.10762037602072623, 0.3783985793590546, 0.3688957737551795, 0.3822804285420312, 0.39233422146903146, 0.18592197238293587, 0.39835565818680657, 0.36646339164839853, 0.29996280868848163, 0.35952309105131364, 0.19767125672509347, 0.3922528200679355, 0.3683128794034322, 0.2780546369890838, 0.3555277672078874, 0.3687988817691803]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.5316045413682643
Per Topic: [0.5580196784602272, 0.6070075160927243, 0.3320080794477474, 0.19004649233998394, 0.5593037558926477, 0.625640649927987, 0.5640980223814647, 0.5553434246116215, 0.5791655408011542, 0.6549664517243703, 0.6875664088461134, 0.6306196133295695, 0.532685252360555, 0.6239841527409024, 0.1635345952412056, 0.6113076812691158, 0.5893126898341708, 0.6207689881324768, 0.6512066079510583, 0.26687552570191325, 0.6637576699256897, 0.5820300175084008, 0.4329821487267812, 0.5680274128913879, 0.29369887570285214, 0.6470509966214498, 0.5842362410492368, 0.4285337212423949, 0.557279497385025, 0.5870785329076978]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.NONE: 0> =S===
Average: 0.5844497149251596
Per Topic: [0.5936046262814404, 0.6503789352872539, 0.35685782028959667, 1.5368331864793847, 0.5886256409357892, 0.5971336365555293, 0.5579301101135489, 1.0312153663442274, -0.015801263126963696, 0.8471407045467768, 0.5532432510658585, 0.5334630190756529, 1.0494534604589918, 0.9489622076240566, -1.1309658679251207, 0.6300692686554132, 0.5668640611007268, 0.48795462530517447, 0.6444204216480794, 1.1198775447601903, 0.6320095574554566, 0.5382207289094766, 1.1684190336359297, 0.7448662141188256, -1.003586996406117, 0.5806891532275694, 0.74467025670524, 0.9061688786978715, 0.6163566360309041, 0.458417229904026]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.11935449063389972
Per Topic: [0.09104066548170522, 0.06104483656365321, 0.3093050652080112, 0.28175778368280996, 0.08212855363470023, 0.008025267011382514, 0.12080522253006873, 0.1865225060661841, 0.030552612939532587, 0.1300905320340664, 0.059253980701517626, 0.041128822967099646, -0.013953080879420869, 0.09569137090713614, 0.17172210053023365, 0.13141210740238118, 0.09200627624264193, 0.09485439706542012, 0.03239099182925808, 0.3284517904122671, 0.029139023562897312, 0.09320585189982214, 0.20915542462219794, 0.14141440639117112, 0.19086722037431578, 0.07778486348171201, 0.1352714304941603, 0.24153346777929982, 0.08869124026872062, 0.039339987812046374]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.DICE: 2> =S===
Average: 9.764055218641406
Per Topic: [1.4785268584887186, 8.029365785916646, 0.9702214002609253, 1.8041651871469286, 6.375461806191338, -0.7050536228550806, 1.4947402609719171, 2.4129498998324075, 3.9956230812602573, 5.715696952078077, -3.1508254329363505, -3.0207321961720783, -0.4067401882674959, 1.2127181026670668, 1.1320184853341844, 211.9972156604131, 1.2562444779607984, 1.7112665004200405, -0.5146351165241665, 0.9359890752368503, -4.61125920481152, 5.525151567988925, 39.07899330457052, 1.171814935737186, 0.9255641990237766, 2.0271761390897964, 1.5980050881703696, 1.2008336848682828, 1.1204772697554695, 2.160682597425249]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.JACCARD: 3> =S===
Average: -20.825088196727965
Per Topic: [-2.7075881520907084, -2.272716283135944, 3.2274680429034763, 8.821884269184537, -4.040207725101047, -2.5036019414663313, -0.882104041841295, 18.635549034012687, -1.0613703462812636, -3.0882180836465625, -1.289473260111279, -2.3885018050670626, -2.2382670289940303, -2.2146502597464455, -0.25286320050557454, -6.070304331514571, -3.884779371155633, -2.4781847609413994, -2.379622534248564, -604.0762025992076, -1.6511220229996575, -0.15254030227661133, -5.448412187894186, -3.880208917458852, 4.612057495117187, -3.180569440788693, -3.9969701210657758, -0.6741435077455309, 6.395788433154424, -3.6327709509266746]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.NONE: 0> =S===
Average: 0.2978416590213352
Per Topic: [0.3780543494180079, 0.4129695691956162, 0.18980654629698515, 0.7691373178155086, 0.4330623607524941, 0.30900114777448207, 0.355847175203072, 0.5755005634088165, -0.2125246165440904, 0.36983900020384447, 0.28614394288966066, 0.3204577164651876, 0.2715209622206863, 0.5262686189508876, -1.4235381631689017, 0.362400284509572, 0.37013832469472235, 0.28775052264550244, 0.35168678990791136, 0.8617703727076175, 0.2960053444307032, 0.31365885996615045, 0.8398320378929996, 0.4166274363483794, -0.9223808370094702, 0.3299575706066556, 0.48694471889678254, 0.6346445910731939, 0.42487595515039467, 0.3197913079366835]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.17994302767120876
Per Topic: [0.12805647044959995, 0.11670115184194098, 0.32550658281478617, 0.3205361928480367, 0.1428330622962676, 0.17555796836192408, 0.16637061934840555, 0.2242692696241041, 0.06496669764682236, 0.2734027441797985, 0.23776455167163577, 0.1406386292203226, 0.10335524227573639, 0.2027769971690658, 0.1911078766376401, 0.19030399062256848, 0.15196374676500757, 0.13059370080526506, 0.12559629707700676, 0.3441504937079218, 0.1384310316397912, 0.1333530227472592, 0.21875141391323671, 0.17642224971204995, 0.20602956772264508, 0.14699623237570955, 0.16957532687520144, 0.2709827545409401, 0.10145181030060889, 0.07984513494496544]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.DICE: 2> =S===
Average: 2.7841405712333676
Per Topic: [1.666050275001261, 0.6319933338297739, 0.9500268008973863, 1.0386896597014532, 2.250735392794013, 0.2756188680003915, 0.17857569058736164, 0.9083827204174466, -2.425627151255806, 0.8707619605792893, -0.2592729784548283, -0.19907247605216172, 0.802290514152911, 0.8496030672556825, 1.077572684817844, 1.2845817714929582, 3.02141981518103, 1.3063109305169847, -0.03761112979716725, 0.9118721087773641, -0.33338644554217656, 1.1887123200628493, 1.0431044512324863, 1.445079586075412, 0.9786576999558343, 61.388298433356816, 1.0510625998179117, 1.1705422348446317, 1.6549388404418197, -1.1656944416877297]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.JACCARD: 3> =S===
Average: 7.0628127242821375
Per Topic: [-9.166540891263221, -2.342065739466084, -6.15212869114346, -43.87733449406094, 56.44401636720738, -2.6669036716222765, 1.8036339667108323, 0.8181499573919508, -3.2101934199945794, -3.841389800608158, -1.0703579567372798, -2.519921213326355, -4.007283565733168, -2.817425860464573, -3.0735806226730347, -1.983945485121674, -5.066192412128051, -8.380187131961186, -2.8874885131087566, -4.33076155450609, -1.4770257814062966, -2.8912580755021837, 289.9129321826829, -3.833444407582283, -8.179676280419033, -4.677306281195746, -2.820743430985345, 2.5057655387454565, -4.367011500419014, -3.959949502845605]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.NONE: 0> =S===
Average: 0.1766937747568545
Per Topic: [0.14824682586212107, 0.18708438146379353, 0.20281956968249237, 0.2442528284296655, 0.1741918908510146, 0.14265440707643534, 0.14944404510495854, 0.25219999875725335, 0.10834079577106827, 0.2163301201716119, 0.19242175789761945, 0.1524210869864975, 0.15412284717444538, 0.247149721569528, 0.111412935305373, 0.17405880215656605, 0.15648258888424163, 0.1349883197180439, 0.18527317950462321, 0.24086584898178115, 0.15769056713778107, 0.1293967353327041, 0.25091810534198355, 0.18260028518696708, 0.12661959749425447, 0.1588389619937186, 0.1887080040766936, 0.2339521853385075, 0.16486529111142117, 0.13246155834246956]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.30799426429340176
Per Topic: [0.2577278188119332, 0.36455040506811603, 0.27655373497141733, 0.37316074293080925, 0.30523637334505715, 0.31315051799433097, 0.25819794668091667, 0.3904047669635879, 0.226201621008416, 0.4634987852639622, 0.43473703985412915, 0.3618507874508699, 0.286240578111675, 0.5096372005012301, 0.16657824232760404, 0.31665071715186865, 0.29665989984447755, 0.26487721796664926, 0.34197703417804504, 0.3128788981690175, 0.3643739732810193, 0.22974988263514307, 0.2334528881435593, 0.34796334778269133, 0.1884039598180809, 0.3390489612188604, 0.2937325588117043, 0.29894404822132653, 0.18890943971152108, 0.2344785405840311]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.DICE: 2> =S===
Average: -6.696977050953278
Per Topic: [-0.501081607511474, -0.058775322056478924, -5.950174129340383, -1.0146991021827692, -0.19650192408718997, -0.5972039008306133, 68.61222103497842, -0.5676436835382548, -0.43249382376670836, -0.2505337163909442, 0.004097158461809159, -0.45448160244462393, 2.5815661487686965, 0.10029054445525011, 3.3434767007827757, -12.810671669674209, -0.8501870264609654, -0.34551202019469607, 13.165061561514934, 0.3700122098128001, -0.013202836861213048, 0.8229714737584194, -0.018339197296235295, -0.1298573174048215, -0.5295765637523598, -270.0585608153294, -0.7114471904726491, 5.680398739212089, 0.04282868843939569, -0.14129233918696021]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.JACCARD: 3> =S===
Average: 2.1071468107418276
Per Topic: [-0.2321656532999542, -0.009382950059241718, -1.7773478090763093, -0.3373492727159626, -0.12773477082244225, -0.14153379789657064, -0.58121748691968, -0.5889425377361477, -0.20039150988062224, -0.03989723877360423, 0.0805715151131153, -0.027872106507937942, -0.122680844925344, 0.12404205043696695, -1.1835544238487878, -0.5044889359683212, -0.24931146192053955, -0.21346006609706414, -0.1409227941185236, 9.27308953884575, 0.0325390866647164, -0.3664802716837989, -0.6178307306435373, -0.08279434137253297, 63.20406588969959, -0.19907079886438117, -0.30779310458650194, -0.6767237674030993, -0.5674934085458517, -0.2034636748385512]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.NONE: 0> =S===
Average: -1.2919665389931534
Per Topic: [-1.17386418557316, -0.8680202687611864, -1.9011274512865044, -1.047667736559903, -0.8366067935289703, -1.0824157815953828, -1.0848464209397464, -1.042256352533198, -2.040907684969454, -0.9648597265277666, -0.6370809895484615, -1.0602086759498055, -0.9300058656023976, -0.8755223027370869, -3.8756595693460896, -0.9561534903362163, -1.0182683509407169, -1.0452619935102012, -0.8683999688876635, -1.6542543949045063, -0.9976812325252717, -1.3517324017228247, -1.4340397599847348, -1.1336003900525302, -3.3037643601664413, -0.9452589069885053, -1.3000326329631926, -1.1188280272579738, -1.1003419572424118, -1.1103284968522937]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.4489587392133696
Per Topic: [0.4172645108567344, 0.3680855777528551, 0.47551360925038655, 0.46617349055078294, 0.42326092951827576, 0.6414405763149261, 0.38452149331569674, 0.4236492501364814, 0.45179317759142984, 0.4734817637337579, 0.513688063621521, 0.4954831686284807, 0.5364161756303575, 0.4674039426777098, 0.3921919425328573, 0.45504247943560283, 0.40978858404689367, 0.34588676674498453, 0.4799913730886247, 0.5026300244861179, 0.48717684911357034, 0.3698685281806522, 0.45938059820069205, 0.48731302552753025, 0.452699175145891, 0.40655995971626707, 0.4119377543528875, 0.42373811536365086, 0.4009397059679031, 0.44544156491756437]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.DICE: 2> =S===
Average: 0.7170875298093867
Per Topic: [0.7426078796386719, 0.7015290180842082, 0.7337837298711141, 0.7339953780174255, 0.7182757245169745, 0.7243074602550931, 0.7232087903552585, 0.7609699527422588, 0.7437359650929769, 0.704906948407491, 0.6405402686860826, 0.6788614100880093, 0.7135660039054023, 0.6997373965051439, 0.7442686306105719, 0.7021779139836629, 0.7368649337026808, 0.714368814892239, 0.682016827000512, 0.7392514573203193, 0.7062499708599514, 0.7357237325774298, 0.7286302831437853, 0.7034077525138855, 0.7649813916948106, 0.6894622564315795, 0.6908912089135911, 0.7162305646472507, 0.7261502756012811, 0.7119239542219374]
==== Coherence : <Segmentation.ONE_PRE: 2>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.JACCARD: 3> =S===
Average: 2.9189544194716004
Per Topic: [3.515099125438266, 2.5720635890960692, 3.0092657830980087, 3.2807204246521, 2.874766953786214, 3.7289518859651354, 2.8617791864607067, 3.7924955818388195, 3.4049933248096043, 2.7408284081353083, 1.896372185813056, 2.4040002981821695, 3.1123066107432047, 2.618385820918613, 3.2927805900573732, 2.509185202916463, 3.3346821069717407, 2.619035180409749, 2.6232512129677668, 3.124668386247423, 2.6490831957923042, 3.215196074379815, 2.9361532661649914, 2.5999829795625473, 3.845041184955173, 2.3929948170979816, 2.3027957253985933, 2.6045406103134154, 2.966953155729506, 2.740259716245863]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.NONE: 0> =S===
Average: 0.041999999999236134
Per Topic: [0.04666666666620373, 0.12999999999861117, 0.03666666666620372, 0.03999999999930559, 0.028888888887808717, 0.19888888888711426, -0.007777777778395056, 0.13777777777700623, 0.0011111111108024763, 0.1388888888878087, 0.028888888887808717, -0.06666666666689813, 0.014444444443209914, -0.06888888888919752, 0.018888888888503125, 0.15333333333240745, 0.028888888887808717, 0.05999999999930558, 0.10666666666550928, 0.02111111111080248, 0.0755555555547068, -0.013333333333796271, 0.05666666666620372, -0.08222222222229938, 0.0011111111108024763, 0.05555555555401248, 0.016666666666203728, 0.15333333333240745, -0.026666666667592542, -0.0244444444452932]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.15673434886460502
Per Topic: [0.28987571597099304, 0.320284903049469, 0.15148071944713593, 0.20374327898025513, 0.08384250849485397, 0.34609341621398926, 0.01871933601796627, 0.4937671720981598, 0.026877814903855324, 0.40721428394317627, 0.08384250849485397, -0.19972720742225647, 0.2210555225610733, -0.22205013036727905, 0.008636128157377243, 0.40278589725494385, 0.3336375951766968, 0.23292849957942963, 0.35347050428390503, 0.26747798919677734, 0.3050496578216553, -0.13561280071735382, 0.14727157354354858, -0.10859449952840805, 0.014918585307896137, 0.25913405418395996, 0.05369899049401283, 0.5455716252326965, 0.013203505426645279, -0.21656668186187744]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.DICE: 2> =S===
Average: -0.3440898985369131
Per Topic: [-0.44132310152053833, -0.1838964819908142, -0.7713922262191772, -0.15017744898796082, -0.13535699248313904, -0.0145052969455719, -0.5916098952293396, -0.9177692532539368, -0.38130316138267517, -0.014815865084528923, -0.13535699248313904, -1.1045511960983276, -0.03331481292843819, -0.3744139075279236, -0.5134019255638123, -0.002610116498544812, -0.0871577113866806, -0.32402369379997253, -0.04791494831442833, -0.4453229010105133, -0.577428936958313, -0.6193571090698242, -0.283419132232666, -0.317392498254776, -0.6215577125549316, -0.00806919950991869, -0.57356196641922, -0.04842308908700943, -0.2178465873003006, -0.38542279601097107]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.22468823567808915
Per Topic: [-0.3061930537223816, -0.155331552028656, -0.4354722201824188, -0.1305689513683319, -0.11921977251768112, -0.014297901652753353, -0.3717052936553955, -0.47856080532073975, -0.2760459780693054, -0.014599558897316456, -0.11921977251768112, -0.5248393416404724, -0.032240718603134155, -0.2724171280860901, -0.3392370343208313, -0.0026033215690404177, -0.08017025142908096, -0.2447265088558197, -0.04572408273816109, -0.30811312794685364, -0.36605697870254517, -0.3824710249900818, -0.22083134949207306, -0.24092480540275574, -0.383309006690979, -0.008004610426723957, -0.36449912190437317, -0.04618659242987633, -0.178878515958786, -0.2781986892223358]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.NONE: 0> =S===
Average: 1.1829920572622323
Per Topic: [1.3888888888567388, 1.351351351341206, 1.2820512820238879, 1.1904761904604435, 1.0802469135719117, 1.4520202020110353, 0.9661835748675581, 1.9841269840876163, 1.0101010100755026, 1.5555555555382716, 1.0802469135719117, 0.5555555555452675, 1.033591731259473, 0.617283950607758, 1.1574074073806158, 1.8518518518232738, 1.0802469135719117, 1.315789473664974, 1.3440860214933326, 1.234567901196464, 1.3285024154428924, 0.9259259259116369, 1.5151515151132535, 0.25252525251887564, 1.0101010100755026, 1.1111111111049383, 1.111111111090535, 1.8518518518232738, 0.9259259259187814, 0.9259259259181319]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.8453894654909769
Per Topic: [0.8197667598724365, 0.9256532788276672, 0.8035576939582825, 0.8818578124046326, 0.8918190002441406, 0.9353445768356323, 0.8346521258354187, 0.873748779296875, 0.7586309909820557, 0.9246693849563599, 0.8918190002441406, 0.7219855189323425, 0.9121947288513184, 0.7525374889373779, 0.7535561919212341, 0.8993445038795471, 0.9197915196418762, 0.8803811073303223, 0.9235726594924927, 0.7482755780220032, 0.8938221335411072, 0.8005030751228333, 0.8128131628036499, 0.7002859115600586, 0.7354151606559753, 0.9233746528625488, 0.805790901184082, 0.9214084148406982, 0.8755013942718506, 0.8396104574203491]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.DICE: 2> =S===
Average: 0.38887887398401894
Per Topic: [0.3785381019115448, 0.42045146226882935, 0.3751106262207031, 0.3976234495639801, 0.4064190685749054, 0.42867404222488403, 0.3846781253814697, 0.3993554711341858, 0.35036173462867737, 0.42043137550354004, 0.4064190685749054, 0.34244710206985474, 0.415459543466568, 0.34844744205474854, 0.36161744594573975, 0.40465158224105835, 0.41982099413871765, 0.39895099401474, 0.4213864803314209, 0.3619428873062134, 0.4089191257953644, 0.37124642729759216, 0.37123751640319824, 0.32440507411956787, 0.3589421510696411, 0.4214225709438324, 0.3721025586128235, 0.41501474380493164, 0.3991887867450714, 0.38110026717185974]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.6396074990431467
Per Topic: [0.609109103679657, 0.7254809141159058, 0.6002833247184753, 0.6600911617279053, 0.6846902370452881, 0.7503141760826111, 0.6251657009124756, 0.6648780703544617, 0.5393182039260864, 0.7254211902618408, 0.6846902370452881, 0.520790159702301, 0.7107454538345337, 0.5347955226898193, 0.5664587616920471, 0.6796886920928955, 0.7236059308052063, 0.6637577414512634, 0.7282691597938538, 0.5672577619552612, 0.6918159127235413, 0.5904481410980225, 0.5904256105422974, 0.4801769256591797, 0.5599215030670166, 0.7283771634101868, 0.5926167368888855, 0.7094447016716003, 0.6644162535667419, 0.6157705187797546]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.NONE: 0> =S===
Average: 1.513354111031174
Per Topic: [1.7777777777432102, 1.6842105263022469, 1.5238095237835223, 1.3333333333185187, 1.1313131313048466, 1.947089947075267, 0.9481481481391936, 4.444444444320989, 1.0158730158556815, 2.2626262625994085, 1.1313131313048466, 0.44444444443950626, 1.053497942380003, 0.5079365079314689, 1.269841269819602, 3.5555555554864204, 1.1313131313048466, 1.5999999999795558, 1.6666666666516203, 1.4222222221898273, 1.6296296296115227, 0.8888888888790125, 2.133333333284741, 0.17777777777550618, 1.0158730158556815, 1.1851851851779427, 1.1851851851687243, 3.5555555554864204, 0.8888888888827162, 0.8888888888823795]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.756882642582059
Per Topic: [0.7650192379951477, 0.8613271713256836, 0.7436153888702393, 0.8194134831428528, 0.828204870223999, 0.8931920528411865, 0.7570669054985046, 0.84662264585495, 0.6689490675926208, 0.8754242658615112, 0.828204870223999, 0.639968752861023, 0.862912118434906, 0.6384710669517517, 0.05105343088507652, 0.7766585350036621, 0.8765435218811035, 0.8180068135261536, 0.879450798034668, 0.7161286473274231, 0.8418636918067932, 0.7145481109619141, 0.7389515042304993, 0.5211626291275024, 0.689755380153656, 0.8764177560806274, 0.7430278062820435, 0.870980441570282, 0.8075026869773865, 0.7560356259346008]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.DICE: 2> =S===
Average: 0.3428099542940226
Per Topic: [0.34372544288635254, 0.38872024416923523, 0.34047603607177734, 0.3650163412094116, 0.37595123052597046, 0.4036736488342285, 0.34979259967803955, 0.3731122314929962, 0.3054954707622528, 0.3928699791431427, 0.37595123052597046, 0.30285680294036865, 0.386983186006546, 0.30006176233291626, 2.5203056308598093e-10, 0.3609868586063385, 0.3934319317340851, 0.3656192421913147, 0.3954794108867645, 0.32332706451416016, 0.3795117437839508, 0.3331964910030365, 0.3284977972507477, 0.24949350953102112, 0.32146450877189636, 0.39533209800720215, 0.33701446652412415, 0.38405901193618774, 0.3673328459262848, 0.34486544132232666]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.535612565287408
Per Topic: [0.5237525105476379, 0.635912299156189, 0.5162451863288879, 0.57484370470047, 0.6024388074874878, 0.6769340634346008, 0.5379707217216492, 0.5951818227767944, 0.43987542390823364, 0.6470937132835388, 0.6024388074874878, 0.4344255328178406, 0.6312766671180725, 0.4286973774433136, 2.5203056308598093e-10, 0.5649130344390869, 0.6486195921897888, 0.57634037733078, 0.654203474521637, 0.4778187870979309, 0.6116340756416321, 0.4996921420097351, 0.48919835686683655, 0.33243343234062195, 0.47376230359077454, 0.6538003087043762, 0.5083285570144653, 0.6235321164131165, 0.5806098580360413, 0.5264039039611816]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.NONE: 0> =S===
Average: 0.24354935751654835
Per Topic: [0.57536414488468, 0.521296923625839, 0.4212134650598962, 0.2876820724414198, 0.1233790211540655, 0.6663359231022626, -0.0532445145272019, 1.4916548767501645, 0.01574835695205985, 0.8165262017090233, 0.1233790211540655, -0.8109302162251897, 0.05211600113348184, -0.677398823599758, 0.23889190826607307, 1.268511325444344, 0.1233790211540655, 0.47000362923358285, 0.5108256237575629, 0.3522205935672775, 0.4883527679034347, -0.1177830356663694, 0.7576857016752074, -1.7272209480976366, 0.01574835695205985, 0.16989903679013044, 0.1698990367823524, 1.268511325444344, -0.11778303566220273, -0.11778303566258155]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.10306088653936361
Per Topic: [0.04774315282702446, 0.00027956522535532713, 0.1759595423936844, 0.044969458132982254, 0.010081577114760876, -0.02322317101061344, 0.25080054998397827, 0.16333380341529846, 0.155797079205513, 0.02492894046008587, 0.010081577114760876, 0.14169538021087646, 0.031085219234228134, 0.2477717101573944, 0.16646547615528107, 0.28673261404037476, 0.030864188447594643, 0.012386607937514782, -0.016495274379849434, 0.16911812126636505, 0.25782379508018494, 0.22052617371082306, 0.20629671216011047, 0.17674943804740906, 0.002177574671804905, 0.022487830370664597, 0.025386424735188484, 0.2607489824295044, 0.019804658368229866, -0.030551111325621605]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.DICE: 2> =S===
Average: 1.6876316348711649
Per Topic: [1.1808829307556152, 3.1761984825134277, 1.11403226852417, 1.6925550699234009, 7.197402000427246, -44.1919059753418, 1.2981277704238892, 1.0437235832214355, 1.3059754371643066, 3.1600587368011475, 7.197402000427246, 1.1293100118637085, -6.238952159881592, 1.9343199729919434, 1.3272883892059326, 2.058518886566162, 2.8631575107574463, 1.5730520486831665, -8.661764144897461, 1.0858306884765625, 1.2325001955032349, 1.4062464237213135, 1.4607523679733276, 1.4230997562408447, 1.191633939743042, 6.938048362731934, 1.2817714214324951, 1.389438509941101, 48.58103942871094, 4.479205131530762]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.JACCARD: 3> =S===
Average: -4.202050316333771
Per Topic: [-6.528441429138184, -1.4595166444778442, -9.76945686340332, -2.4439284801483154, -1.1613579988479614, -0.9778721332550049, -4.3542680740356445, -23.871009826660156, -4.2682366371154785, -1.4629502296447754, -1.1613579988479614, -8.733351707458496, -0.8618583679199219, -2.0702972412109375, -4.055410385131836, -1.9447168111801147, -1.5367234945297241, -2.7450411319732666, -0.8964992761611938, -12.650808334350586, -5.301071643829346, -3.4615578651428223, -3.170363187789917, -3.3635103702545166, -6.2182841300964355, -1.168405532836914, -4.548975944519043, -3.5677988529205322, -1.0210167169570923, -1.2874221801757812]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.NONE: 0> =S===
Average: 0.10650250803827695
Per Topic: [0.3285040669655546, 0.30110509278196973, 0.2484613592937987, 0.17435338714266152, 0.07718963869055703, 0.37295582954090467, -0.03440142671690969, 0.6851790109009271, 0.010050335853248927, 0.44183275227507113, 0.07718963869055703, -0.5877866648873042, 0.033039854077990205, -0.48242614923472477, 0.14618251017493344, 0.6161861394167182, 0.07718963869055703, 0.2744368456982516, 0.2957142441467513, 0.21072103130978853, 0.2840523043982159, -0.07696104113489391, 0.41551544395307993, -1.3762440251916417, 0.010050335853248927, 0.1053605156572708, 0.10536051565597443, 0.6161861394167182, -0.07696104113551105, -0.07696104113545482]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.16347259391720095
Per Topic: [0.07990583777427673, 0.036092985421419144, 0.2575892210006714, 0.07601101696491241, 0.024034013971686363, 0.03873555362224579, 0.36558184027671814, 0.2416311651468277, 0.21751579642295837, 0.065586157143116, 0.024034013971686363, 0.18903659284114838, 0.1921483874320984, 0.34645339846611023, 0.24603603780269623, 0.4383622705936432, 0.05610727518796921, 0.03076649084687233, -0.01406999584287405, 0.23927003145217896, 0.3885394036769867, 0.318610280752182, 0.3091985285282135, 0.22767148911952972, 0.004244009032845497, 0.05279702693223953, 0.04347703978419304, 0.4040156900882721, 0.07553057372570038, -0.07073431462049484]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.DICE: 2> =S===
Average: 1.2288841644922892
Per Topic: [1.1499894857406616, 8.694544792175293, 1.0880317687988281, 1.6410962343215942, -7.84513521194458, -1.5090053081512451, 1.2635118961334229, 1.0226049423217773, 1.2583255767822266, 3.92311429977417, -7.84513521194458, 1.1067636013031006, -1.1698986291885376, 1.8490943908691406, 1.211580514907837, 1.8340578079223633, 4.200595378875732, 1.5275579690933228, -1.2552475929260254, 1.0609850883483887, 1.174046516418457, 1.382623314857483, 1.387900710105896, 1.328208327293396, 1.1547659635543823, -7.241245746612549, 1.2499561309814453, 1.2993130683898926, -2.234434127807617, 24.157958984375]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.JACCARD: 3> =S===
Average: -5.430280049641927
Per Topic: [-7.667129993438721, -1.1299620866775513, -12.359515190124512, -2.559828281402588, -0.8869431614875793, -0.6014356017112732, -4.794895172119141, -45.238197326660156, -4.871084690093994, -1.3421005010604858, -0.8869431614875793, -10.366501808166504, -0.5391488671302795, -2.177724838256836, -5.726335525512695, -2.1989564895629883, -1.3124418258666992, -2.8955254554748535, -0.556589663028717, -17.397472381591797, -6.74558687210083, -3.6135365962982178, -3.5779788494110107, -4.046844959259033, -7.461368560791016, -0.8786590695381165, -5.00070333480835, -4.340982913970947, -0.6908267736434937, -1.0431815385818481]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.NONE: 0> =S===
Average: 0.06624470394148856
Per Topic: [0.11676363675255017, 0.17559212082902823, 0.08831321989894252, 0.07240752719538407, 0.03926002438804054, 0.2537673862402816, -0.013620396947472472, 0.29756946355149305, 0.0031223123848645212, 0.2247240034286678, 0.03926002438804054, -0.16762497072759863, 0.01802913252790635, -0.14987410993346564, 0.048796920694947676, 0.2906177127978444, 0.03926002438804054, 0.11397136410118747, 0.15587535132395897, 0.06546416910010913, 0.12868913731679366, -0.02735506788144294, 0.14769100063500679, -0.2988475929530144, 0.0031223123848645212, 0.06546416910128068, 0.03744939017575551, 0.2906177127978444, -0.03629786572368363, -0.034866993991498695]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.1795986601151526
Per Topic: [0.2393397092819214, 0.2939641773700714, 0.1676352471113205, 0.23427294194698334, 0.1200600191950798, 0.45229148864746094, 0.08748222142457962, 0.39698004722595215, 0.004797083325684071, 0.4234808087348938, 0.1200600191950798, -0.13449855148792267, 0.2900974750518799, -0.16396880149841309, 0.033560678362846375, 0.5256941318511963, 0.32555586099624634, 0.1784573495388031, 0.3900874853134155, 0.31562480330467224, 0.3369654417037964, -0.0987323746085167, 0.1808038353919983, -0.1088065505027771, 0.01916022039949894, 0.2978954613208771, 0.07305517047643661, 0.5543202757835388, 0.03412267193198204, -0.20179854333400726]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.DICE: 2> =S===
Average: -35.873382854368536
Per Topic: [-2.413581609725952, -0.19152449071407318, -7.951733589172363, -0.2445598989725113, -0.12968221306800842, 0.028639836236834526, -1.0611329078674316, 3.067133665084839, -0.9410110116004944, -0.00830205250531435, -0.12968221306800842, -1076.655029296875, -0.011695269495248795, -0.49732834100723267, -1.262964129447937, 0.006568645592778921, -0.09923006594181061, -0.47071945667266846, -0.006059555336833, 22.188892364501953, -0.884681761264801, -1.017217755317688, -0.44068923592567444, -0.7247352600097656, -4.022817134857178, -0.005134203936904669, -1.5642764568328857, -0.15523123741149902, -0.20906424522399902, -0.39463675022125244]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.39027324233514565
Per Topic: [-0.7070525884628296, -0.1607390195131302, -0.8882899284362793, -0.19650308787822723, -0.11479531228542328, 0.029484260827302933, -0.5148299932479858, -1.4837617874145508, -0.48480454087257385, -0.008233696222305298, -0.11479531228542328, -0.999072253704071, -0.011560071259737015, -0.33214378356933594, -0.5581017136573792, 0.006612077821046114, -0.09027235209941864, -0.32006070017814636, -0.006023058667778969, -1.0471946001052856, -0.46940648555755615, -0.5042677521705627, -0.30588778853416443, -0.4202008843421936, -0.8009085059165955, -0.005107979290187359, -0.6100264191627502, -0.13437245786190033, -0.17291411757469177, -0.2829674184322357]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.NONE: 0> =S===
Average: -1.4736030129460684
Per Topic: [-1.7917594692248326, -0.693147180560723, -1.7917594692248326, -1.3862943611186682, -0.9444616088410577, -0.4480247225281728, -1.5040773967745518, -1.280933845461242, -2.197224577329997, -0.9444616088410577, -0.9444616088410577, -2.484906649778778, -0.8109302162168566, -2.197224577329997, -1.9740810260175872, -1.0986122886678875, -0.9444616088410577, -1.3862943611186682, -0.8754687373542778, -2.197224577329997, -1.1856236656572445, -1.7917594692248326, -1.7917594692248326, -3.5835189384228876, -2.197224577329997, -0.5877866649030966, -1.7917594692248326, -1.0986122886678875, -1.0986122886678875, -1.1856236656572445]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.4051099667946498
Per Topic: [0.3128114342689514, 0.29022714495658875, 0.456012099981308, 0.27219223976135254, 0.27618899941444397, 0.5187976956367493, 0.5151218771934509, 0.45409396290779114, 0.4310186803340912, 0.2844853401184082, 0.27618899941444397, 0.406873881816864, 0.5763556361198425, 0.5211324095726013, 0.44940024614334106, 0.5777320265769958, 0.25635379552841187, 0.26241374015808105, 0.5060144066810608, 0.4632687270641327, 0.5369794368743896, 0.5030347108840942, 0.4982015788555145, 0.4322194457054138, 0.26809704303741455, 0.2195909172296524, 0.2773457169532776, 0.5402519106864929, 0.5302295088768005, 0.24066539108753204]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.DICE: 2> =S===
Average: 0.7169607857863108
Per Topic: [0.8160318732261658, 0.6320065855979919, 0.7978934049606323, 0.7204073071479797, 0.6506289839744568, 0.6404455900192261, 0.7013998627662659, 0.8310660719871521, 0.7743260264396667, 0.6533805131912231, 0.6506289839744568, 0.800041675567627, 0.6519516706466675, 0.7004050612449646, 0.7702807784080505, 0.6710249781608582, 0.6511701941490173, 0.714690089225769, 0.6047012209892273, 0.8539513349533081, 0.680045485496521, 0.7045339941978455, 0.741414487361908, 0.7823833227157593, 0.8255860805511475, 0.7111220359802246, 0.7788773775100708, 0.6950854659080505, 0.6371696591377258, 0.6661734580993652]
==== Coherence : <Segmentation.ONE_SUC: 3>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.JACCARD: 3> =S===
Average: 2.79475519657135
Per Topic: [4.435728073120117, 1.7174402475357056, 3.947885513305664, 2.5766310691833496, 1.8622870445251465, 1.7812200784683228, 2.3489603996276855, 4.919475555419922, 3.431169271469116, 1.885008454322815, 1.8622870445251465, 4.001043796539307, 1.8731648921966553, 2.3378403186798096, 3.353139877319336, 2.0397439002990723, 1.866727590560913, 2.5049610137939453, 1.5297322273254395, 5.847029685974121, 2.125443935394287, 2.3844850063323975, 2.8671932220458984, 3.595236301422119, 4.7334885597229, 2.461669683456421, 3.522379159927368, 2.279606819152832, 1.7561087608337402, 1.9955683946609497]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.NONE: 0> =S===
Average: 0.05055555552748148
Per Topic: [-0.253, 0.6079999999150001, -0.13799999999999998, 0.7759999999074999, -0.25999999999999995, -0.33999999999999997, -0.24899999999999997, -0.26, -0.26599999999999996, -0.343, 0.4843333332497222, -0.29000000000000004, -0.367, 0.50299999993, -0.14900000000000002, 0.7029999999000001, 0.6749999999074999, -0.296, -0.332, -0.10500000000000002, 0.6379999999500001, 0.7369999999000001, -0.13699999999999998, 0.613999999915, -0.145, 0.5583333332497221, -0.21400000000000002, -0.177, -0.21000000000000002, -0.24900000000000003]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.5415724128422639
Per Topic: [0.5373274594545364, 0.6175571918487549, 0.4770723730325699, 0.6316824167966842, 0.5644107758998871, 0.5595276564359665, 0.5154653616249562, 0.6185740262269974, 0.5269137233495712, 0.6371875941753388, 0.5716313332319259, 0.5969623416662216, 0.5010495139285922, 0.73782097697258, 0.3656056709587574, 0.5169796407222748, 0.5813835948705673, 0.4918918490409851, 0.5391562134027481, 0.5122612968087197, 0.5736709266901017, 0.4945665955543518, 0.5182306855916977, 0.5902392476797104, 0.42227031141519544, 0.54073486328125, 0.5509150475263596, 0.510292598605156, 0.43703860640525816, 0.508752492070198]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.DICE: 2> =S===
Average: -0.05027374318626243
Per Topic: [0.03927813754417002, 0.04180443454533815, -0.2468165174126625, 0.027459732769057155, 0.025011882465332747, 0.0523207476362586, -0.15124760381877422, -0.7090898901224136, 0.04631739230826497, -0.04348974861204624, 0.033568563591688874, 0.05140705183148384, 0.07163909580558539, 0.03393064513802528, -0.17329260855913162, -0.15716274566948413, 0.031568868551403284, -0.023677445948123932, 0.03465194166637957, -0.35570975244045255, 0.04182005319744349, 0.024973296327516437, 0.0360838046297431, 0.023332816548645496, -0.1388972859829664, -0.06800140934064984, -0.02806585220969282, -0.09312676191329956, 0.019143888726830482, 0.04605297315865755]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.029941071485324454
Per Topic: [0.04308585688704625, 0.0446071389131248, -0.1967119961977005, 0.02873232145793736, 0.02626017779111862, 0.056498335860669616, -0.12840020135045052, -0.41183939278125764, 0.04950031088665128, -0.040606487262994054, 0.03527016192674637, 0.05498626958578825, 0.08003168953582644, 0.03558255136013031, -0.14648936986923217, -0.13402051515877247, 0.03433660753071308, -0.02260943448636681, 0.037681630812585354, -0.2607411935925484, 0.04467113250866532, 0.027608619537204504, 0.037783720856532456, 0.026072914712131023, -0.11964958906173706, -0.06285469895228743, -0.02633865411626175, -0.08349631540477276, 0.022510839439928532, 0.05030542407184839]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.NONE: 0> =S===
Average: 1.3552114872225542
Per Topic: [0.0, 3.754637752793726, 0.0, 8.385281375197048, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.2441218301929142, 0.0, 0.0, 2.9333105725176685, 0.0, 3.8905305203853535, 4.541123934836883, 0.0, 0.0, 0.0, 3.17222144025044, 4.939386063134093, 0.0, 3.8182960730456528, 0.0, 2.9774350543228523, 0.0, 0.0, 0.0, 0.0]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.9146278921763101
Per Topic: [0.9166006147861481, 0.9460621774196625, 0.8346863150596618, 0.8896758079528808, 0.9241811692714691, 0.937481838464737, 0.9202969610691071, 0.9123888492584229, 0.9169389188289643, 0.9522718071937561, 0.9645551145076752, 0.9446840107440948, 0.9442847371101379, 0.9547281682491302, 0.8270976185798645, 0.9409907221794128, 0.9267397880554199, 0.9392213821411133, 0.9492060959339141, 0.8096889436244965, 0.9583303928375244, 0.9172113299369812, 0.8633925139904022, 0.9372662842273712, 0.8148745715618133, 0.9502478659152984, 0.9243027687072753, 0.8960786700248718, 0.9017829835414887, 0.9235683441162109]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.DICE: 2> =S===
Average: 0.09997679015000661
Per Topic: [0.09999999850988388, 0.10000000223517418, 0.09983752146363259, 0.09997299909591675, 0.09999999776482582, 0.09999999850988388, 0.1, 0.09998270869255066, 0.10000000074505806, 0.10000000298023223, 0.09999999701976776, 0.09999999403953552, 0.10000000670552253, 0.1, 0.09982802346348763, 0.10000000149011612, 0.10000000521540642, 0.10000000223517418, 0.10000000223517418, 0.09972630068659782, 0.1000000037252903, 0.09999999627470971, 0.09997882321476936, 0.10000000223517418, 0.09999999850988388, 0.10000000149011612, 0.10000000447034836, 0.09997731372714043, 0.09999999776482582, 0.1]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.1111955966303746
Per Topic: [0.11120680719614029, 0.11128147840499877, 0.11113375052809715, 0.11124621704220772, 0.1112453892827034, 0.11116116121411324, 0.11123602688312531, 0.1111360877752304, 0.11117456331849099, 0.11113841906189918, 0.11113927066326142, 0.11117525771260262, 0.11122014597058297, 0.11120561137795448, 0.11129894480109215, 0.1112654909491539, 0.11122472137212754, 0.11114718317985535, 0.11117018386721611, 0.11086780652403831, 0.1111651249229908, 0.11118221059441566, 0.11119996532797813, 0.11132487207651139, 0.11122125610709191, 0.11115459278225899, 0.111176498234272, 0.11129455417394638, 0.11122925654053688, 0.1112450510263443]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.NONE: 0> =S===
Average: 1.4393458787279667
Per Topic: [0.0, 3.91612019971099, 0.0, 9.500013296804193, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.2771944399682726, 0.0, 0.0, 3.023488366399701, 0.0, 4.031166594645543, 4.80990493381229, 0.0, 0.0, 0.0, 3.3473098207303265, 5.222771860531334, 0.0, 3.992185453036204, 0.0, 3.0602213962001406, 0.0, 0.0, 0.0, 0.0]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.703916822627378
Per Topic: [0.872867226600647, 0.9013971149921417, 0.0466287637129426, 0.11647273153066635, 0.871793395280838, 0.914996188879013, 0.888783973455429, 0.8472237527370453, 0.8838164329528808, 0.92010298371315, 0.9369237124919891, 0.9197418034076691, 0.05875986196674117, 0.9177050948143005, 0.07719137948006391, 0.8989042401313782, 0.8865294039249421, 0.9135325968265533, 0.9234955966472626, 0.10785636231303215, 0.9338362574577331, 0.8839708745479584, 0.7614807307720184, 0.8858444094657898, 0.0600609116256237, 0.9261069238185883, 0.8955683052539826, 0.10792532227933407, 0.8649173259735108, 0.8930710017681122]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.DICE: 2> =S===
Average: 0.07714912438359542
Per Topic: [0.10000000596046447, 0.09977788031101227, 3.2780742120763763e-09, 0.01643266061982747, 0.09973965883255005, 0.1, 0.09999999627470971, 0.09973392114043236, 0.09999999776482582, 0.09999999701976776, 0.09999999850988388, 0.09999999776482582, 2.743425109308362e-09, 0.09995957762002945, 1.5768337677246613e-10, 0.09989505112171174, 0.09971871227025986, 0.09999999776482582, 0.09999999925494193, 7.87408849234339e-10, 0.09999999701976776, 0.10000000074505806, 0.09936678409576416, 0.09984949454665185, 2.908799512368887e-09, 0.09999999478459358, 0.09999999701976776, 1.9358580072292853e-09, 0.1, 0.09999999925494193]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.08589823905593837
Per Topic: [0.11150938868522645, 0.11153265833854675, 3.2780742120763763e-09, 0.01723609397849675, 0.11136268749833107, 0.11126035824418068, 0.11148354113101959, 0.11087922155857086, 0.11130479648709297, 0.11118842512369156, 0.11112882122397423, 0.11124122738838196, 2.743425109308362e-09, 0.11147894263267517, 1.5768337538468736e-10, 0.11138427332043647, 0.11109846010804177, 0.11119341179728508, 0.11115589812397957, 7.87408849234339e-10, 0.11124403774738312, 0.11130044609308243, 0.11093106865882874, 0.1117858462035656, 2.908799512368887e-09, 0.11121246218681335, 0.11125476285815239, 1.9358580072292853e-09, 0.11129547506570817, 0.11148485541343689]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.NONE: 0> =S===
Average: -17.998075495427734
Per Topic: [-27.631021115928547, 1.2140641365107059, -27.631021115928547, 1.99268501437722, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, 0.7556502297443957, -27.631021115928547, -27.631021115928547, 1.0629795470195091, -27.631021115928547, 1.316640100991759, 1.4100873610433413, -27.631021115928547, -27.631021115928547, -27.631021115928547, 1.1340019846351324, 1.5058705645390194, -27.631021115928547, 1.2574427025423358, -27.631021115928547, 1.0287358143354586, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.2581733147753403
Per Topic: [0.24508911790326238, 0.17374768014997244, 0.5105727970600128, 0.48035172671079635, 0.2207817789982073, 0.058474055491387844, 0.2861169584095478, 0.377871535718441, 0.1149206195026636, 0.27650429494678974, 0.1526490034069866, 0.12525155174080282, -0.02420790456235409, 0.2585714322398417, 0.34213744625449183, 0.3058826974593103, 0.23692641034722328, 0.2380435297265649, 0.11151184756308793, 0.530694967508316, 0.09301305944100022, 0.23980206837877632, 0.3962100613862276, 0.33828441947698595, 0.38168871533125637, 0.19976201755926012, 0.3081607738509774, 0.43378524780273436, 0.21452752351760865, 0.11807400994002819]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.DICE: 2> =S===
Average: 1.4955023588736858
Per Topic: [1.2561096668243408, 1.8136528134346008, 0.9747179567813873, 1.2463554620742798, 1.4402464628219604, 1.3493381023406983, 1.1479509949684144, 1.0095133185386658, 1.325226652622223, 1.5603938579559327, 4.509990119934082, 2.0536433339118956, 1.7466441154479981, 2.458484387397766, 0.9938374519348144, 1.266217577457428, 1.2540597200393677, 1.3152145147323608, 1.4414571046829223, 0.9607468128204346, 2.6165847301483156, 1.1989105820655823, 1.0787838399410248, 1.52461838722229, 0.9657190799713135, 1.5096396565437318, 1.245335590839386, 1.0928990066051483, 1.0921587586402892, 1.416620707511902]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.JACCARD: 3> =S===
Average: -7.092051823536552
Per Topic: [-6.871484327316284, -2.254611086845398, -49.44212474822998, -7.2994259119033815, -3.3631428241729737, -4.071650528907776, -11.374973249435424, -13.510870265960694, -4.225879836082458, -2.798841881752014, -1.2853970527648926, -1.9529073238372803, -2.342919707298279, -1.6996152639389037, -45.50078015327453, -5.234318804740906, -5.3679927587509155, -4.312158226966858, -3.308485984802246, 1.496824312210083, -1.6213171839714051, -6.712603521347046, -0.5913944244384766, -3.05693256855011, 7.437892961502075, -3.0004308223724365, -5.7403112888336185, -6.814890384674072, -14.517410898208619, -3.423400950431824]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.NONE: 0> =S===
Average: -0.7321758026569418
Per Topic: [-2.1193269466817144, 1.1836213392910842, 0.0, 1.9093945089752844, -2.05001222868822, 0.0, 0.0, -6.282835324932625, -2.09055873945737, -2.159873457471698, 0.7431858331334158, -2.1821878125947856, -6.468966903707212, 1.035352912605204, 0.0, 1.2868936721207613, 1.3698277353097963, 0.0, -2.272474583729167, 0.0, 1.0883911517599014, 1.4648907739423103, -2.151172319776523, 1.2238002725221477, 0.0, 1.0063391125479737, -2.1716517610327064, -2.163955656922057, -2.163955656922057, 0.0]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.3125091246325367
Per Topic: [0.2762430664151907, 0.25201512277126314, 0.5245177417993545, 0.5173569798469544, 0.28970395103096963, 0.09718930795788765, 0.34388906210660936, 0.4112602941691875, 0.12733350574271754, 0.4111373096704483, 0.3453836627304554, 0.22875212784856558, 0.017550497874617578, 0.36196450125426055, 0.36375135481357573, 0.37084027417004106, 0.287821763753891, 0.2903820521198213, 0.1693682083627209, 0.5441355347633362, 0.21100983675569296, 0.27975747734308243, 0.4058368131518364, 0.3833363808691502, 0.39377264669165013, 0.29187180995941164, 0.3566843113861978, 0.4650069460272789, 0.2167186163365841, 0.1406825812533498]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.DICE: 2> =S===
Average: 3.8570949314037963
Per Topic: [1.2249209702014923, 1.882071053981781, 0.9637438952922821, 1.1076842308044434, 1.405919623374939, 1.3376509308815003, 1.1359732747077942, 0.9947576940059661, 1.3089823722839355, 1.5321281790733337, 74.32208251953125, 2.260452628135681, 1.8589930772781371, 2.706422758102417, 0.9735818028450012, 1.2448540449142456, 1.2417418599128722, 1.2851081848144532, 1.444409430027008, 0.944037276506424, 3.541761589050293, 1.1833403348922729, 1.0524930655956268, 1.4987923622131347, 0.9549902737140655, 1.5251638889312744, 1.2416791915893555, 1.0531478464603423, 1.0766732692718506, 1.4092903137207031]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.JACCARD: 3> =S===
Average: 20.507366928259525
Per Topic: [7.635204005241394, -2.1603544473648073, 742.0231639385223, -2.4880732536315917, -3.6107279777526857, -4.309712243080139, -49.40712885856628, -30.79481234550476, -4.448127388954163, -2.899846577644348, -1.013667368888855, -1.797537386417389, -2.1667372226715087, -1.601905071735382, 27.565062618255617, -6.282110285758972, -5.812558126449585, -4.776027750968933, -3.317580962181091, -22.963042879104613, -1.3958378553390502, -7.697141504287719, -9.76264271736145, -3.2256082773208616, 44.59204630851745, -2.959061336517334, -6.136319494247436, -0.9429188728332519, -21.148436403274538, -3.476552414894104]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.NONE: 0> =S===
Average: 0.04916006557067793
Per Topic: [-0.07670099985772799, 0.2570201081641083, 0.0, 0.41461974952117087, -0.0741924165627901, 0.0, 0.0, -0.22738339269375524, -0.07565984372007947, -0.07816842701577136, 0.16138075318276104, -0.07897601045720357, -0.23411971915790059, 0.22482402838833196, 0.0, 0.27944541030519776, 0.2974543133079955, 0.0, -0.08224359766491388, 0.0, 0.27821696095948567, 0.318096989863972, -0.07785352234183011, 0.26574485265980285, 0.0, 0.21852376175624544, -0.07859469803599864, -0.0783161667403813, -0.0783161667403813, 0.0]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.4864306248476107
Per Topic: [0.43763422667980195, 0.5497331604361534, 0.47610455825924874, 0.5632155805826187, 0.4909617304801941, 0.46946394741535186, 0.45332501977682116, 0.5710992306470871, 0.38939291536808013, 0.6348734751343728, 0.6118693023920059, 0.541075611114502, 0.45656153485178946, 0.6805806457996368, 0.3075068242847919, 0.5102350264787674, 0.4688612878322601, 0.4481358230113983, 0.5128358319401741, 0.5088331699371338, 0.5522102773189544, 0.40120257437229156, 0.41351083368062974, 0.5533533722162247, 0.35195610448718073, 0.5260767102241516, 0.48553305864334106, 0.48568335771560667, 0.3347507417201996, 0.40634281262755395]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.DICE: 2> =S===
Average: 0.13213661321516457
Per Topic: [-0.2333679087460041, -0.07421420821920037, 2.4051308393478394, -0.6659291565418244, -0.1743314877152443, -0.128420652449131, -1.008475089073181, 4.273856472969055, -0.18330565243959426, -0.17203156054019927, 0.0013614120427519083, -0.04500359818339348, -0.015463637164793908, -0.06803527064621448, 2.363974618911743, -0.5117599874734878, -0.26517586410045624, -0.24317131489515303, -0.11499179378151894, 1.4837201237678528, -0.02400804781354964, -0.3645357459783554, -1.1503982663154602, -0.1648890905082226, 1.8324528217315674, -0.242745341360569, -0.40612800121307374, -1.3859741806983947, -0.6302479207515717, -0.12379411570727825]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.5107786150136963
Per Topic: [-0.18751066848635672, -0.06765787499025464, -1.7275664448738097, -0.3980284184217453, -0.14733463600277902, -0.1131181314587593, -0.49960836470127107, -1.3078935027122498, -0.15389785915613174, -0.146271950006485, 0.0017750745406374334, -0.0421593233011663, -0.012756952526979148, -0.06273228777572512, -1.7361850976943969, -0.3371551722288132, -0.20854200422763824, -0.1951710104942322, -0.10176142491400242, -3.1382298469543457, -0.022615865571424365, -0.26592340022325517, -0.5338767528533935, -0.13852907791733743, -2.2275139331817626, -0.1945817530155182, -0.28789310455322265, -0.580046021938324, -0.3826148182153702, -0.10795782655477523]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.NONE: 0> =S===
Average: -18.455730713128055
Per Topic: [-27.631021115928547, -0.1386294362007891, -27.631021115928547, -0.06931471814989455, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -0.17917594700983888, -27.631021115928547, -27.631021115928547, -0.2772588723025781, -27.631021115928547, -9.900003040065772e-11, -0.06931471814989455, -27.631021115928547, -27.631021115928547, -27.631021115928547, -4.8999915237636707e-11, -9.900003040065772e-11, -27.631021115928547, -0.13862943620078907, -27.631021115928547, -0.17917594700983888, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.6004041439294816
Per Topic: [0.578860592842102, 0.5383680820465088, 0.6549134254455566, 0.6475066512823104, 0.5820991456508636, 0.6632340461015701, 0.5645034700632096, 0.6035295695066452, 0.5770901381969452, 0.6192097634077072, 0.630240973830223, 0.6226463794708252, 0.6046907782554627, 0.6189156651496888, 0.5785669237375259, 0.6159706115722656, 0.5710591435432434, 0.5234188318252564, 0.591490375995636, 0.6765992105007171, 0.6013480097055435, 0.5419433057308197, 0.6310879856348037, 0.6410084575414657, 0.6261020600795746, 0.5683780044317246, 0.591088530421257, 0.6096499741077424, 0.5589282512664795, 0.5796759605407715]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.DICE: 2> =S===
Average: 0.9081074988842013
Per Topic: [0.9051498711109162, 0.9094338595867157, 0.9018328309059143, 0.9027936637401581, 0.9086619794368744, 0.9168468117713928, 0.9072303771972656, 0.9084529221057892, 0.9122909307479858, 0.9121033608913421, 0.9074244916439056, 0.9111976206302643, 0.9152511358261108, 0.905062073469162, 0.9079863071441651, 0.9058874368667602, 0.9098460018634796, 0.9077514350414276, 0.9143188178539277, 0.9023345887660981, 0.9079844713211059, 0.9084756791591644, 0.903959184885025, 0.9047776937484742, 0.9054386794567109, 0.9087440133094787, 0.9028993308544159, 0.9061625897884369, 0.9105643033981323, 0.9123625040054322]
==== Coherence : <Segmentation.ONE_ALL: 4>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.JACCARD: 3> =S===
Average: 13.783973473707835
Per Topic: [16.25237317085266, 12.821036100387573, 12.807693099975586, 14.896856880187988, 13.771287155151366, 18.443945384025575, 13.258090591430664, 17.12075710296631, 16.368441152572633, 13.711332082748413, 10.349943590164184, 12.753678321838379, 15.601322603225707, 13.092324829101562, 13.680219697952271, 12.417581510543823, 15.442327070236207, 12.039814376831055, 13.387025356292725, 13.295765495300293, 13.442171192169189, 14.691886854171752, 12.82756757736206, 12.966483163833619, 16.276408004760743, 12.039440584182739, 11.163235473632813, 11.66354718208313, 13.309474611282349, 13.62717399597168]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.NONE: 0> =S===
Average: 0.07499999996833334
Per Topic: [-0.253, 0.7079999999, -0.13799999999999998, 0.8259999998999998, -0.25999999999999995, -0.33999999999999997, -0.24899999999999997, -0.26, -0.26599999999999996, -0.343, 0.6009999999, -0.29000000000000004, -0.367, 0.7029999999000001, -0.14900000000000002, 0.7029999999000001, 0.7249999999, -0.296, -0.332, -0.10500000000000002, 0.6379999999500001, 0.7369999999000001, -0.13699999999999998, 0.7139999999, -0.145, 0.6749999999, -0.21400000000000002, -0.177, -0.21000000000000002, -0.24900000000000003]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.6410729912131234
Per Topic: [0.6333910822868347, 0.6973669797182083, 0.5970727115869522, 0.7139965176582337, 0.6602881044149399, 0.6524575412273407, 0.6169759377837181, 0.6953143775463104, 0.6318665146827698, 0.7100220680236816, 0.6676555216312409, 0.6833051443099976, 0.6037357442080975, 0.7908996045589447, 0.5168040946125985, 0.6219531387090683, 0.6696984320878983, 0.6087498277425766, 0.6363498419523239, 0.6194979637861252, 0.6657921642065048, 0.6044882446527481, 0.6306056916713715, 0.6719021423435378, 0.5522374004125595, 0.6434971481561661, 0.6501025140285492, 0.6179495543241501, 0.555858987569809, 0.6123547405004501]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.DICE: 2> =S===
Average: -0.04950055712210693
Per Topic: [0.03739694920368493, 0.03908383771777153, -0.23679817765951156, 0.026073369197547436, 0.025373382633551954, 0.050565505400300026, -0.1477519493550062, -0.7227683544158936, 0.047584916590858485, -0.04714265819638967, 0.03233479708433151, 0.04880514014512301, 0.0674881803803146, 0.03012196607887745, -0.1613907128572464, -0.15321893766522407, 0.029766265489161014, -0.02233187196252402, 0.031501518096774817, -0.34623068273067475, 0.038427666574716565, 0.029923668410629035, 0.04161144886165857, 0.02365320026874542, -0.13267237772233784, -0.06120393336750567, -0.029163892567157745, -0.09083370342850686, 0.023013164289295673, 0.04376556184142828]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.DIFFERENCE: 1>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.029246959519878154
Per Topic: [0.040554215712472794, 0.041367638669908044, -0.18859166279435158, 0.027406617300584912, 0.026466488419100643, 0.05443309042602777, -0.12601786963641642, -0.4160135120153427, 0.0507568078530312, -0.0440474362578243, 0.03395622838288546, 0.05208738949149847, 0.07485087784007191, 0.031389358546584846, -0.13628447167575358, -0.13091368600726128, 0.03229295266792178, -0.02133885601942893, 0.03413750594481826, -0.25280910134315493, 0.04088243441656232, 0.0326758299022913, 0.043880258593708275, 0.025354486890137196, -0.11289855823852121, -0.056442486308515075, -0.027269362611696123, -0.08109336085617543, 0.026242190320044757, 0.04757720679044723]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.NONE: 0> =S===
Average: 1.4820239801682076
Per Topic: [0.0, 4.143526641437244, 0.0, 8.839826829122668, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.5244248604016497, 0.0, 0.0, 4.073038898727946, 0.0, 3.8905305203853535, 4.957790600982717, 0.0, 0.0, 0.0, 3.17222144025044, 4.939386063134093, 0.0, 4.439508193554606, 0.0, 3.4804653570495008, 0.0, 0.0, 0.0, 0.0]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.9317025850216545
Per Topic: [0.9332154810428619, 0.9568761706352233, 0.8678891599178314, 0.9118741154670715, 0.939766138792038, 0.9502161026000977, 0.9357312977313995, 0.9302503347396851, 0.9338732779026031, 0.9616746664047241, 0.9712950348854065, 0.9553639352321625, 0.9548657298088074, 0.96359983086586, 0.8633557140827179, 0.9525385618209838, 0.9414067029953003, 0.9512531757354736, 0.9588610708713532, 0.8481564521789551, 0.9664158940315246, 0.933635026216507, 0.8913208246231079, 0.9496804177761078, 0.8528951346874237, 0.9599951684474946, 0.9392126202583313, 0.916973489522934, 0.9204763114452362, 0.9384097099304199]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.DICE: 2> =S===
Average: 0.09084771491587165
Per Topic: [0.090857744961977, 0.09081744477152824, 0.09078982919454574, 0.09081626906991005, 0.09083664193749427, 0.09088162183761597, 0.09084171429276466, 0.09088353589177131, 0.09087421372532845, 0.09089420288801194, 0.0908934824168682, 0.090873484313488, 0.0908483773469925, 0.09085730984807014, 0.09069855883717537, 0.09082573726773262, 0.09084807336330414, 0.09088974744081497, 0.09087632820010186, 0.09085912853479386, 0.09087948873639107, 0.09087031185626984, 0.09084642007946968, 0.0907941222190857, 0.0908469520509243, 0.09088497683405876, 0.090873122215271, 0.09079249650239944, 0.09084264189004898, 0.09083746895194053]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.RATIO: 2>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.10000000116725759
Per Topic: [0.09999999925494193, 0.10000000521540642, 0.09999999329447747, 0.1000000037252903, 0.09999999850988388, 0.1000000074505806, 0.10000000819563866, 0.09999999925494193, 0.09999999329447747, 0.10000000521540642, 0.10000000521540642, 0.10000000670552253, 0.10000001266598701, 0.09999999478459358, 0.09999999403953552, 0.10000000223517418, 0.10000000298023223, 0.1000000037252903, 0.09999999925494193, 0.10000000447034836, 0.10000001266598701, 0.09999999552965164, 0.09999999105930328, 0.09999999478459358, 0.09999998658895493, 0.10000000447034836, 0.10000000447034836, 0.10000000447034836, 0.10000000074505806, 0.10000000074505806]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.NONE: 0> =S===
Average: 1.574614330098895
Per Topic: [0.0, 4.322135237242685, 0.0, 10.000013296724692, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 2.56441368063523, 0.0, 0.0, 4.235554731218934, 0.0, 4.031166594645543, 5.264450388285885, 0.0, 0.0, 0.0, 3.3473098207303265, 5.222771860531334, 0.0, 4.664599246033868, 0.0, 3.58601504691836, 0.0, 0.0, 0.0, 0.0]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.7202866797702516
Per Topic: [0.8995069801807404, 0.9238621056079864, 0.03974424544721842, 0.11647273153066635, 0.9006467580795288, 0.9329968214035034, 0.9107627511024475, 0.8805878221988678, 0.9084170997142792, 0.9366410970687866, 0.9493901491165161, 0.9355605125427247, 0.05246623436214254, 0.9351759076118469, 0.07719137948006391, 0.9200849056243896, 0.9109231293201446, 0.9309146285057068, 0.9388108253479004, 0.10785636231303215, 0.94701087474823, 0.9077460765838623, 0.8153495192527771, 0.9106474757194519, 0.0600609116256237, 0.9409192383289338, 0.9164608538150787, 0.09641246907413006, 0.8911628544330596, 0.9148176729679107]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.DICE: 2> =S===
Average: 0.08373983587039549
Per Topic: [0.09070029854774475, 0.090534558147192, 0.05000000283627202, 0.068759411375146, 0.09059121534228325, 0.09082677140831948, 0.09071093946695327, 0.09085014089941978, 0.09080255255103112, 0.09086695313453674, 0.09089943617582322, 0.09083719700574874, 0.050000002527070084, 0.09068491533398629, 0.07616382919974851, 0.09069322496652603, 0.09071683958172798, 0.09086524546146393, 0.0908846378326416, 0.06374088437782353, 0.09083639308810235, 0.0908079169690609, 0.09055597260594368, 0.09045128077268601, 0.06621104702854534, 0.09085262939333916, 0.09082993492484093, 0.050000001652027934, 0.0908067338168621, 0.09071410968899726]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LIKELIHOOD: 3>, <IndirectMeasure.JACCARD: 3> =S===
Average: 0.09999999938123459
Per Topic: [0.10000000894069672, 0.10000000149011612, 0.10000000283627201, 0.09999999831161985, 0.10000000223517418, 0.09999999403953552, 0.09999999105930328, 0.09999998137354851, 0.09999998882412911, 0.10000000298023223, 0.10000000223517418, 0.10000000596046447, 0.10000000252707009, 0.09999999925494193, 0.09999999635644484, 0.10000000521540642, 0.09999999403953552, 0.09999999180436134, 0.10000000074505806, 0.10000000359045096, 0.09999999701976776, 0.10000000968575477, 0.10000001043081283, 0.09999999403953552, 0.10000000820064922, 0.09999999925494193, 0.09999998435378074, 0.10000000165202796, 0.10000000074505806, 0.10000000223517418]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.NONE: 0> =S===
Average: -17.962483372280577
Per Topic: [-27.631021115928547, 1.3547240469054427, -27.631021115928547, 2.063014969574606, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, 0.9378823009649313, -27.631021115928547, -27.631021115928547, 1.3442993678090223, -27.631021115928547, 1.316640100991759, 1.4804173162407261, -27.631021115928547, -27.631021115928547, -27.631021115928547, 1.1340019846351324, 1.5058705645390194, -27.631021115928547, 1.3981026129370882, -27.631021115928547, 1.2109678855560513, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.4232210506374636
Per Topic: [0.40471704192459584, 0.3748432844877243, 0.6054804116487503, 0.577781380712986, 0.39561948031187055, 0.2525869905948639, 0.4440085932612419, 0.49328284449875354, 0.3226379733532667, 0.44289947897195814, 0.3710936218500137, 0.3367249697446823, 0.23194885086268185, 0.4129734568297863, 0.4842275597155094, 0.450000374764204, 0.3989551968872547, 0.4163795210421085, 0.3081002600491047, 0.6187533557415008, 0.3349212914705276, 0.4033030778169632, 0.5193201501853764, 0.4590431611053646, 0.4998456843197346, 0.39013231322169306, 0.4592802584171295, 0.5574072420597076, 0.39356791619211434, 0.33679577708244324]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.DICE: 2> =S===
Average: 1.4709807678063713
Per Topic: [1.2335412502288818, 1.7493083596229553, 0.9758998453617096, 1.242317545413971, 1.3921151995658874, 1.305328941345215, 1.1145889222621919, 0.9970690608024597, 1.267557430267334, 1.5003574252128602, 4.813571095466614, 2.008640003204346, 1.6858244955539703, 2.4432090878486634, 0.975768095254898, 1.230519461631775, 1.2183526396751403, 1.2709275364875794, 1.3993108928203584, 0.9570511102676391, 2.5273105978965758, 1.1639196932315827, 1.0658064484596252, 1.497915780544281, 0.9561408758163452, 1.4438908696174622, 1.2263261437416078, 1.0679799139499664, 1.0561725080013276, 1.342701804637909]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGLIKELIHOOD: 4>, <IndirectMeasure.JACCARD: 3> =S===
Average: -8.774326799710591
Per Topic: [-180.41927905082701, -2.694600486755371, -8.551981115341187, 91.20622534751892, -5.142442607879639, -1.0554541110992433, -2.1000454425811768, -15.980204629898072, -6.960401701927185, -28.277136898040773, -1.3747661471366883, -18.568219244480133, -1.4437849283218385, -1.7968761801719666, -127.09476642608642, -12.824231719970703, 5.664210915565491, -6.707182431221009, -2.5311867952346803, -0.9237553596496582, -1.7390028595924378, -7.049778604507447, -6.394920063018799, -3.4542298316955566, -3.7749561309814452, -22.10694603919983, -6.989784550666809, -18.917453050613403, 144.33578805923463, -9.562641906738282]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.NONE: 0> =S===
Average: 0.4454398795410999
Per Topic: [0.0, 1.3222507753641843, 0.0, 1.9787092269858242, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.9223617800281911, 0.0, 0.0, 1.3126117847152092, 0.0, 1.2868936721207613, 1.439142453324124, 0.0, 0.0, 0.0, 1.0883911517599014, 1.4648907739423103, 0.0, 1.3624297085720154, 0.0, 1.1855150594204762, 0.0, 0.0, 0.0, 0.0]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.44621886669153665
Per Topic: [0.4124165214598179, 0.41815333664417265, 0.6163956195116043, 0.6051172554492951, 0.4346828691661358, 0.23162247641012074, 0.4828745499253273, 0.5160080606117845, 0.3062569569796324, 0.5360476672649384, 0.47579583078622817, 0.3852004334330559, 0.1935215424746275, 0.4705633774399757, 0.49999310076236725, 0.4825298238545656, 0.4240686010569334, 0.4432110205292702, 0.31062971577048304, 0.629476198554039, 0.38552218191325666, 0.4223848611116409, 0.5205207884311676, 0.4717846617102623, 0.506538450345397, 0.4375795140862465, 0.4865640806034207, 0.580752494931221, 0.37728322446346285, 0.32307078506564724]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.DICE: 2> =S===
Average: 1.0435703825950624
Per Topic: [1.2180174231529235, 1.8631192445755005, 0.9675707817077637, 1.1141510665416718, 1.378911566734314, 1.3160255074501037, 1.1139546632766724, 0.9873750627040863, 1.2584400296211242, 1.502124309539795, -10.115685820579529, 2.321642255783081, 1.88246488571167, 2.8269158482551573, 0.9587222039699554, 1.229571372270584, 1.2208441019058227, 1.2584603667259215, 1.4352289259433746, 0.9435745537281036, 3.802091729640961, 1.1612120151519776, 1.045456701517105, 1.5008124589920044, 0.946215808391571, 1.4917181432247162, 1.2405412077903748, 1.0362650513648988, 1.0487590491771699, 1.3526109635829926]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.PMI: 5>, <IndirectMeasure.JACCARD: 3> =S===
Average: -11.746772848367693
Per Topic: [5.979167127609253, -2.6410277843475343, 26.23333306312561, -1.7838737487792968, -8.457412910461425, -1.4535639047622682, -3.577716016769409, -27.5216344833374, -8.907718539237976, -0.33805098533630373, -1.0935552358627318, 0.4839033007621765, -1.4398733735084535, -1.685006892681122, 49.72143075466156, -0.885801887512207, 0.6011057376861573, -27.695952582359315, -2.4876408100128176, -20.633373832702638, -1.4742779970169066, -9.314956903457642, -11.319663143157959, -3.6779964208602904, 28.433544445037843, -0.4355814456939697, -7.089989948272705, -313.0866086959839, -11.304293155670166, 4.449900817871094]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.NONE: 0> =S===
Average: 0.09812190169672579
Per Topic: [0.0, 0.28712310772271554, 0.0, 0.4296712492948265, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.2002883156967014, 0.0, 0.0, 0.2850300274976867, 0.0, 0.27944541030519776, 0.3125058130824736, 0.0, 0.0, 0.0, 0.27821696095948567, 0.318096989863972, 0.0, 0.29584785221336524, 0.0, 0.25743132426534937, 0.0, 0.0, 0.0, 0.0]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.6040872625385721
Per Topic: [0.5692125529050827, 0.65080326795578, 0.5849519714713096, 0.6516749501228333, 0.6060932964086533, 0.5997911095619202, 0.5725445315241814, 0.6604177474975585, 0.5421510398387909, 0.7167915105819702, 0.6984683573246002, 0.6483241111040116, 0.5904821053147316, 0.744459581375122, 0.48266232311725615, 0.6158387660980225, 0.5940763026475906, 0.5786906272172928, 0.6278219133615494, 0.6103616148233414, 0.6522184759378433, 0.5438662767410278, 0.5459191679954529, 0.6375426169484854, 0.5009278669953346, 0.6335693955421448, 0.6010507076978684, 0.6048533856868744, 0.5055979818105698, 0.5514543205499649]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.DICE: 2> =S===
Average: 0.16895263698810598
Per Topic: [-0.22467442005872726, -0.0717879543080926, 2.4869715571403503, -0.6741549015045166, -0.16681005880236627, -0.11651808135211468, -1.0089258432388306, 5.113375616073609, -0.16409189701080323, -0.17005307972431183, 0.0017191542850923725, -0.04115924602374434, -0.01008585398667492, -0.07271983586251736, 2.4648455142974854, -0.5109133243560791, -0.25942555218935015, -0.2349082961678505, -0.11019136048853398, 1.4878202319145202, -0.023804608383215964, -0.35086887776851655, -1.1440933644771576, -0.16545448899269105, 1.8452358961105346, -0.234650719165802, -0.4044821739196777, -1.3970118880271911, -0.6600303232669831, -0.11457271110266447]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.NPMI: 6>, <IndirectMeasure.JACCARD: 3> =S===
Average: -0.53091334684587
Per Topic: [-0.18150741010904312, -0.06580013204365968, -1.7469874143600463, -0.4021384835243225, -0.1423875130712986, -0.10274352580308914, -0.49714879393577577, -1.3276653647422791, -0.14003336653113366, -0.14451706111431123, 0.0022407311684219168, -0.03821370052173734, -0.006127732509048656, -0.0674062106758356, -2.0379212737083434, -0.3348137766122818, -0.20347139984369278, -0.19002137333154678, -0.0963026087731123, -3.2521924495697023, -0.022489816357847303, -0.2565449684858322, -0.5277079284191132, -0.14061422869563103, -2.4841044306755067, -0.1890276476740837, -0.28705792129039764, -0.5744902908802032, -0.37045379281044005, -0.09975052047520876]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.NONE: 0> =S===
Average: -18.420680743983702
Per Topic: [-27.631021115928547, -9.900003040065772e-11, -27.631021115928547, -9.900003040065772e-11, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547, -9.900003040065772e-11, -27.631021115928547, -27.631021115928547, -9.900003040065772e-11, -27.631021115928547, -9.900003040065772e-11, -9.900003040065772e-11, -27.631021115928547, -27.631021115928547, -27.631021115928547, -4.8999915237636707e-11, -9.900003040065772e-11, -27.631021115928547, -9.900003040065772e-11, -27.631021115928547, -9.900003040065772e-11, -27.631021115928547, -27.631021115928547, -27.631021115928547, -27.631021115928547]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.COSINE: 1> =S===
Average: 0.6858331813414893
Per Topic: [0.6675231665372848, 0.6407616078853607, 0.7199262201786041, 0.7087128102779389, 0.674959808588028, 0.7404359757900238, 0.658565005660057, 0.6793989837169647, 0.6806197047233582, 0.7043274402618408, 0.717370119690895, 0.7084757924079895, 0.6948964387178421, 0.6933852344751358, 0.6634315013885498, 0.6951278984546662, 0.6594241172075271, 0.6288093000650405, 0.6775986671447753, 0.7365705847740174, 0.7033872485160828, 0.634419396519661, 0.706954562664032, 0.7053548276424408, 0.7015772461891174, 0.662407511472702, 0.6772064924240112, 0.6904528617858887, 0.6603913962841034, 0.6825235188007355]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.DICE: 2> =S===
Average: 0.9118037780125937
Per Topic: [0.9116515040397644, 0.9112907469272613, 0.9106871962547303, 0.9116369605064392, 0.9115502536296844, 0.9159986495971679, 0.9106798052787781, 0.9134948909282684, 0.911925446987152, 0.9125157237052918, 0.9103051900863648, 0.9119442880153656, 0.9142145872116089, 0.9110994279384613, 0.9133953988552094, 0.9107547461986542, 0.9126778662204742, 0.9105921447277069, 0.9141834139823913, 0.9106616199016571, 0.9110627174377441, 0.9120863258838654, 0.91106196641922, 0.9114029586315155, 0.9120813608169556, 0.9111285269260406, 0.9101901948451996, 0.9104019999504089, 0.9119841039180756, 0.9114533245563508]
==== Coherence : <Segmentation.ONE_SET: 5>, <ConfirmMeasure.LOGCOND: 7>, <IndirectMeasure.JACCARD: 3> =S===
Average: 14.449915270805356
Per Topic: [17.00997748374939, 13.411004638671875, 13.621732091903686, 15.789669513702393, 14.465864181518555, 19.18271050453186, 13.796858215332032, 17.785425448417662, 16.787820911407472, 14.353698372840881, 11.057847213745116, 13.338881254196167, 16.170499539375307, 13.692289924621582, 14.495761942863464, 13.061396408081055, 16.125968885421752, 12.430436897277833, 14.109033560752868, 14.151474952697754, 14.06671895980835, 15.34122953414917, 13.646188926696777, 13.780387353897094, 16.990155601501463, 12.628229808807372, 11.969531393051147, 12.254098796844483, 13.900067400932311, 14.082498407363891]
Pining Topics using Word Priors#
# make LDA model and train
mdl = tp.LDAModel(k=20, min_cf=10, min_df=5, corpus=corpus)
# The word 'church' is assigned to Topic 0 with a weight of 1.0 and to the remaining topics with a weight of 0.1.
# Therefore, a topic related to 'nasdaq' can be fixed at Topic 0 .
mdl.set_word_prior("nasdaq", [1.0 if k == 0 else 0.0001 for k in range(20)])
# Topic 1 for a topic related to 'bank'
mdl.set_word_prior("bank", [1.0 if k == 1 else 0.0001 for k in range(20)])
# Topic 2 for a topic related to 'car'
mdl.set_word_prior("oil", [1.0 if k == 2 else 0.0001 for k in range(20)])
mdl.train(0)
print(
"Num docs:",
len(mdl.docs),
", Vocab size:",
len(mdl.used_vocabs),
", Num words:",
mdl.num_words,
)
print("Removed top words:", mdl.removed_top_words)
for i in range(0, 100, 10):
mdl.train(10)
print("Iteration: {}\tLog-likelihood: {}".format(i, mdl.ll_per_word))
Show code cell output
Num docs: 100 , Vocab size: 726 , Num words: 23802
Removed top words: []
Iteration: 0 Log-likelihood: -6.9982102600171
Iteration: 10 Log-likelihood: -6.78172128538039
Iteration: 20 Log-likelihood: -6.707927669262908
Iteration: 30 Log-likelihood: -6.669206157186698
Iteration: 40 Log-likelihood: -6.661336047859807
Iteration: 50 Log-likelihood: -6.63225589900849
Iteration: 60 Log-likelihood: -6.619234496642365
Iteration: 70 Log-likelihood: -6.6113331038327985
Iteration: 80 Log-likelihood: -6.589239579596395
Iteration: 90 Log-likelihood: -6.604038265559883
mdl.summary()
Show code cell output
<Basic Info>
| LDAModel (current version: 0.12.3)
| 100 docs, 23802 words
| Total Vocabs: 5139, Used Vocabs: 726
| Entropy of words: 6.20784
| Entropy of term-weighted words: 6.20784
| Removed Vocabs: <NA>
|
<Training Info>
| Iterations: 100, Burn-in steps: 0
| Optimization Interval: 10
| Log-likelihood per word: -6.60404
|
<Initial Parameters>
| tw: TermWeight.ONE
| min_cf: 10 (minimum collection frequency of words)
| min_df: 5 (minimum document frequency of words)
| rm_top: 0 (the number of top words to be removed)
| k: 20 (the number of topics between 1 ~ 32767)
| alpha: [0.1] (hyperparameter of Dirichlet distribution for document-topic, given as a single `float` in case of symmetric prior and as a list with length `k` of `float` in case of asymmetric prior.)
| eta: 0.01 (hyperparameter of Dirichlet distribution for topic-word)
| seed: 2936119906 (random seed)
| trained in version 0.12.3
|
<Parameters>
| alpha (Dirichlet prior on the per-document topic distributions)
| [0.91635215 0.40094903 0.46934685 0.31341073 0.8245844 0.24786447
| 0.41051129 0.42268345 0.59450996 0.26436833 0.7942999 0.9315055
| 0.74502105 0.36342382 0.6116812 0.4277962 0.97071016 0.26491374
| 1.0327024 0.6145224 ]
| eta (Dirichlet prior on the per-topic word distribution)
| 0.01
|
<Topics>
| #0 (1420) : nasdaq inc gain global demand
| #1 (870) : bank financi goldman could increas
| #2 (935) : oil data countri product sinc
| #3 (615) : investor fund cash manag activist
| #4 (2113) : zack stock rank compani buy
| #5 (609) : revenu billion report cloud share
| #6 (835) : market invest valu product also
| #7 (1250) : percent index dollar retail yen
| #8 (929) : year sector risk move amazon
| #9 (930) : gold price silver metal see
| #10 (1461) : compani nyse unit time number
| #11 (1492) : price high market last new
| #12 (1752) : year million sale revenu expect
| #13 (714) : dollar market store today rate
| #14 (1257) : quarter compani growth oper earn
| #15 (1027) : said reuter trump govern billion
| #16 (1967) : earn estim stock expect posit
| #17 (467) : peopl googl say media ask
| #18 (1975) : trade week us still forecast
| #19 (1184) : one secur make would presid
|
for k in range(mdl.k):
print("== Topic #{} ==".format(k))
for word, prob in mdl.get_topic_words(k, top_n=10):
print(word, prob, sep="\t")
print()
Show code cell output
== Topic #0 ==
nasdaq 0.07707776129245758
inc 0.0574597492814064
gain 0.04695010185241699
global 0.04624945670366287
demand 0.04134495183825493
recent 0.039243023842573166
world 0.03223659098148346
develop 0.028032733127474785
industri 0.028032733127474785
system 0.025230159983038902
== Topic #1 ==
bank 0.0991838201880455
financi 0.06384652107954025
goldman 0.047887738794088364
could 0.04332808777689934
increas 0.04218817502260208
invest 0.041048262268304825
fed 0.03876843675971031
firm 0.0364886112511158
rise 0.034208785742521286
interest 0.03306887298822403
== Topic #2 ==
oil 0.07005497813224792
data 0.05201324075460434
countri 0.043523017317056656
product 0.03821662440896034
sinc 0.03821662440896034
low 0.03715534880757332
well 0.03715534880757332
declin 0.029726402834057808
energi 0.027603846043348312
rose 0.02654256857931614
== Topic #3 ==
investor 0.12054447084665298
fund 0.08358242362737656
cash 0.05304856225848198
manag 0.046620383858680725
activist 0.04019220173358917
score 0.03537106513977051
flow 0.03537106513977051
even 0.028942884877324104
valu 0.027335839346051216
hedg 0.027335839346051216
== Topic #4 ==
zack 0.08725816011428833
stock 0.08112683892250061
rank 0.0603746697306633
compani 0.037264298647642136
buy 0.036792658269405365
share 0.03396281599998474
current 0.02641656994819641
technolog 0.02594492956995964
hold 0.02547328919172287
nyse 0.024058369919657707
== Topic #5 ==
revenu 0.14605848491191864
billion 0.13632233440876007
report 0.07952811568975449
cloud 0.05681043490767479
share 0.04869697988033295
top 0.042206212878227234
ep 0.042206212878227234
busi 0.032470062375068665
consensu 0.032470062375068665
bp 0.032470062375068665
== Topic #6 ==
market 0.09380713850259781
invest 0.07362334430217743
valu 0.05343955382704735
product 0.047503143548965454
also 0.034443046897649765
money 0.034443046897649765
consum 0.03206848353147507
cost 0.03088119998574257
measur 0.027319354936480522
move 0.027319354936480522
== Topic #7 ==
percent 0.1447671800851822
index 0.05091230198740959
dollar 0.04534463584423065
retail 0.04136773571372032
yen 0.04136773571372032
yield 0.0381862111389637
rate 0.029437027871608734
share 0.027050886303186417
currenc 0.025460125878453255
consum 0.02227860502898693
== Topic #8 ==
year 0.0726400837302208
sector 0.04059769585728645
risk 0.03952961415052414
move 0.03312114253640175
amazon 0.029916903004050255
come 0.02884882315993309
chang 0.027780745178461075
option 0.025644585490226746
good 0.02457650564610958
take 0.023508427664637566
== Topic #9 ==
gold 0.13657896220684052
price 0.05335765704512596
silver 0.050156839191913605
metal 0.03948744013905525
see 0.03842049837112427
season 0.034152742475271225
come 0.032018862664699554
report 0.029884984716773033
mine 0.028818044811487198
investor 0.026684165000915527
== Topic #10 ==
compani 0.10489286482334137
nyse 0.044276900589466095
unit 0.025206705555319786
time 0.02384454943239689
number 0.02316347137093544
work 0.021120237186551094
grow 0.019077003002166748
back 0.019077003002166748
last 0.0183959249407053
deal 0.01771484687924385
== Topic #11 ==
price 0.07671117782592773
high 0.07337620109319687
market 0.06070328131318092
last 0.054033324122428894
new 0.05336632952094078
term 0.05003134906291962
long 0.03935941681265831
see 0.033356454223394394
top 0.02802048809826374
remain 0.026686497032642365
== Topic #12 ==
year 0.14438456296920776
million 0.0716267079114914
sale 0.04491093009710312
revenu 0.03524777293205261
expect 0.03297409042716026
result 0.030131986364722252
increas 0.028995145112276077
billion 0.02672146074473858
per 0.026153041049838066
share 0.02217409573495388
== Topic #13 ==
dollar 0.06379114091396332
market 0.05131297558546066
store 0.04299420490860939
today 0.04160774126648903
rate 0.03467543050646782
buy 0.03328896686434746
open 0.0319025032222271
higher 0.030516041442751884
close 0.030516041442751884
part 0.029129577800631523
== Topic #14 ==
quarter 0.1036258339881897
compani 0.0972980260848999
growth 0.08622435480356216
oper 0.05537626892328262
earn 0.0545852929353714
margin 0.03560185059905052
total 0.03560185059905052
cent 0.03322892263531685
strong 0.0316469706594944
corpor 0.029274040833115578
== Topic #15 ==
said 0.18564963340759277
reuter 0.0435190349817276
trump 0.0425521619617939
govern 0.0328834131360054
billion 0.024181541055440903
york 0.023214666172862053
offici 0.021280916407704353
comment 0.021280916407704353
need 0.019347166642546654
tax 0.019347166642546654
== Topic #16 ==
earn 0.07902201265096664
estim 0.06129385158419609
stock 0.04913739860057831
expect 0.042552653700113297
posit 0.041033096611499786
averag 0.036980945616960526
report 0.033435314893722534
market 0.027357084676623344
also 0.026850566267967224
per 0.023304933682084084
== Topic #17 ==
peopl 0.0717117115855217
googl 0.0506262369453907
say 0.04219204559922218
media 0.040083497762680054
ask 0.037974949926137924
center 0.037974949926137924
would 0.035866402089595795
court 0.035866402089595795
face 0.035866402089595795
new 0.035866402089595795
== Topic #18 ==
trade 0.07214492559432983
week 0.048939090222120285
us 0.027751151472330093
still 0.025228777900338173
forecast 0.022201931104063988
point 0.022201931104063988
time 0.021697456017136574
china 0.021697456017136574
two 0.02119298093020916
one 0.020184030756354332
== Topic #19 ==
one 0.04198075830936432
secur 0.040301863104104996
make 0.03862296789884567
would 0.034425731748342514
presid 0.033586286008358
help 0.030228495597839355
also 0.028549600392580032
gener 0.021834025159478188
two 0.020155129954218864
inc 0.020155129954218864