python - 如何将主题转换为 python 中 LDA 中每个主题中前 20 个单词的列表

Question

我目前正在 python 中研究 LDA 对数。我想将主题隐藏到每个主题中前 20 个单词的列表中。我尝试了下面的代码，但得到了不同的输出。我希望我的输出格式如下:topic=2,words=20。

['men', 'kill', 'soldier', 'order', 'patient', 'night', 'priest', 'becom', 'new', 'speech', 'friend', 'decid', 'young', 'ward', 'state', 'front', 'would', 'home', 'two', 'father']

["n't", 'go', 'fight', 'doe', 'home', 'famili', 'car', 'night', 'say', 'next', 'ask', 'day', 'want', 'show', 'goe', 'friend', 'two', 'polic', 'name', 'meet']

我得到以下输出:

["(u'ngma', 0.034841332255132154)", "(u'video', 0.0073756817356584745)", "(u'youtube', 0.006524039676605746)", "(u'liked', 0.0065240394176856644)",]
["(u'ngma', 0.024537057880333127)", "(u'photography', 0.0068263432438681482)", "(u'tvallwhite', 0.0029535361359022566)", "(u'3', 0.0029252727655122079)"]

我的代码:

`ldamodel = Lda(doc_term_matrix, num_topics=2, id2word = dictionary,passes=50)
lda=ldamodel.print_topics(num_topics=2, num_words=3)

f=open('LDA.txt','w')
f.write(str(lda))
f.close()

topics_matrix = ldamodel.show_topics(formatted=False,num_words=10)
topics_matrix = np.array((topics_matrix),dtype=list)
topic_words = topics_matrix[:, 1]
for i in topic_words:
    print([str(word) for word in i])
    print()`

edit-1:

topic_words = []
for i in range(3):
    tt = ldamodel.get_topic_terms(i,10)
    topic_words.append([pair[0] for pair in tt])
    print topic_words

导致非预期的输出:

[[1897, 135, 130, 127, 70, 162, 445, 656, 608, 1019], [1897, 364, 56, 1236, 181, 172, 449, 48, 15, 18], [1897, 163, 11, 70, 166, 345, 480, 9, 60, 351]]

Answer 1

试试这个-

from gensim import corpora
import gensim
from gensim.models.ldamodel import LdaModel
from gensim.parsing.preprocessing import STOPWORDS

# example docs
doc1 = """
Java (Indonesian: Jawa; Javanese: ꦗꦮ; Sundanese: ᮏᮝ) is an island of Indonesia.\
 With a population of over 141 million (the island itself) or 145 million (the \
 administrative region), Java is home to 56.7 percent of the Indonesian population \
 and is the most populous island on Earth.[1] The Indonesian capital city, Jakarta, \
 is located on western Java. Much of Indonesian history took place on Java. It was \
 the center of powerful Hindu-Buddhist empires, the Islamic sultanates, and the core \
 of the colonial Dutch East Indies. Java was also the center of the Indonesian struggle \
 for independence during the 1930s and 1940s. Java dominates Indonesia politically, \
 economically and culturally.
 """
doc2 = """
Hydrogen fuel is a zero-emission fuel when burned with oxygen, if one considers water \
not to be an emission. It often uses electrochemical cells, or combustion in internal \
engines, to power vehicles and electric devices. It is also used in the propulsion of \
spacecraft and might potentially be mass-produced and commercialized for passenger vehicles \
and aircraft.Hydrogen lies in the first group and first period in the periodic table, i.e. \
it is the first element on the periodic table, making it the lightest element. Since \
hydrogen gas is so light, it rises in the atmosphere and is therefore rarely found in \
its pure form, H2."""

doc3 = """
The giraffe (Giraffa) is a genus of African even-toed ungulate mammals, the tallest living \
terrestrial animals and the largest ruminants. The genus currently consists of one species, \
Giraffa camelopardalis, the type species. Seven other species are extinct, prehistoric \
species known from fossils. Taxonomic classifications of one to eight extant giraffe species\
 have been described, based upon research into the mitochondrial and nuclear DNA, as well \
 as morphological measurements of Giraffa, but the IUCN currently recognizes only one \
 species with nine subspecies.
"""

documents = [doc1, doc2, doc3]
document_wrd_splt = [[word for word in document.lower().split() if word not in STOPWORDS] \
for document in documents]

dictionary = corpora.Dictionary(document_wrd_splt)
print(dictionary.token2id)

corpus = [dictionary.doc2bow(text) for text in texts]

lda = LdaModel(corpus, num_topics=3, id2word = dictionary, passes=50)

num_topics = 3
topic_words = []
for i in range(num_topics):
    tt = lda.get_topic_terms(i,20)
    topic_words.append([dictionary[pair[0]] for pair in tt])

# output
>>> topic_words[0]
['indonesian', 'java', 'species', 'island', 'population', 'million', '(the', 'java.', 'center', 'giraffe', 'currently', 'genus', 'city,', 'economically', 'administrative', 'east', 'sundanese:', 'itself)', 'took', '1940s.']
>>> topic_words[1]
['vehicles', 'fuel', 'hydrogen', 'periodic', 'table,', 'i.e.', 'uses', 'form,', 'considers', 'zero-emission', 'internal', 'period', 'burned', 'cells,', 'rises', 'pure', 'atmosphere', 'aircraft.hydrogen', 'water', 'engines,']
>>> topic_words[2]
['giraffa,', 'even-toed', 'living', 'described,', 'camelopardalis,', 'consists', 'extinct,', 'seven', 'fossils.', 'morphological', 'terrestrial', '(giraffa)', 'dna,', 'mitochondrial', 'nuclear', 'ruminants.', 'classifications', 'species,', 'prehistoric', 'known']