from tmtk.topic_models import anchor, plsa
from tmtk.metrics.metrics import preplexity, coherence, uniq_top_of_topics
from tmtk.collection.collection import FullTextCollection
collection = FullTextCollection(path='./tmtk/corpa/ru_bank_wid_small.zip', lang='ru').fill()
F, anc = anchor.anchor_model(
collection,
wrd_count=len(collection.id_to_words),
metrics=[preplexity, coherence, uniq_top_of_topics])
anchor.print_topics(F, collection.id_to_words, anc, 'ru_bank_wid_small/an.txt')
F, T = plsa.plsa_model(
collection,
wrd_count=len(collection.id_to_words),
metrics=[preplexity, coherence, uniq_top_of_topics],
num_iter=3, verbose=True, F=F)
plsa.print_topics(F, collection.id_to_words, 'ru_bank_wid_small/an+pl.txt')
from tmtk.topic_models import anchor, plsa
from tmtk.metrics.metrics import preplexity, coherence, uniq_top_of_topics
from tmtk.collection.collection import FullTextCollection
from tmtk.collection.transformer_api import TransformerChainApply
from tmtk.collection.transformer import BigramExtractorDocumentsTransform
collection = FullTextCollection(path='./tmtk/corpa/20np.zip', lang='ru').fill()
transformers = TransformerChainApply(
transformers=[BigramExtractorDocumentsTransform(do_apply=False)])
collection = transformers.apply(collection)
F, anc = anchor.anchor_model(
collection,
k=300,
wrd_count=collection.num_wrd,
metrics=[preplexity, coherence, uniq_top_of_topics],
bi=True)
anchor.print_topics(F, collection.id_to_words, anc, 'en_20np/an+bi.txt')
F, T = plsa.plsa_model(collection,
wrd_count=collection.num_wrd,
metrics=[preplexity, coherence, uniq_top_of_topics],
num_iter=3,
verbose=True,
F=F)
plsa.print_topics(F, collection.id_to_words, 'en_20np/an+pl+bi.txt')
from tmtk.topic_models import plsa
from tmtk.metrics.metrics import preplexity, coherence, uniq_top_of_topics
from tmtk.collection.collection import FullTextCollection
collection = FullTextCollection(path='./tmtk/corpa/nips.zip', lang='ru').fill()
F, T = plsa.plsa_model(collection,
wrd_count=len(collection.id_to_words),
metrics=[preplexity, coherence, uniq_top_of_topics],
num_iter=40,
verbose=True)
plsa.print_topics(F, collection.id_to_words, 'en_nips/pl.txt')
print 'F_', F.shape, 'T_', T.shape
from tmtk.topic_models import anchor, plsa
from tmtk.metrics.metrics import preplexity, coherence, uniq_top_of_topics
from tmtk.collection.collection import FullTextCollection
from tmtk.collection.transformer_api import TransformerChainApply
from tmtk.collection.transformer import BigramExtractorDocumentsTransform
collection = FullTextCollection(path='./tmtk/corpa/ru_bank_wid_small.zip', lang='ru').fill()
transformers = TransformerChainApply(transformers=[BigramExtractorDocumentsTransform(do_apply=False)])
collection = transformers.apply(collection)
F, anc = anchor.anchor_model(
collection,
wrd_count=collection.num_wrd,
metrics=[preplexity, coherence, uniq_top_of_topics],
bi=True)
anchor.print_topics(F, collection.id_to_words, anc, 'ru_bank_wid_small/an+bi.txt')
F, T = plsa.plsa_model(
collection,
wrd_count=collection.num_wrd,
metrics=[preplexity, coherence, uniq_top_of_topics],
num_iter=5, verbose=True, F=F)
plsa.print_topics(F, collection.id_to_words, 'ru_bank_wid_small/an+pl+bi.txt')