def lda_terms_analysis(lda_model_filename, word2vec_model_filename):
topics = LDA.get_topics_terms(lda_model_filename)
word2vec = models.Word2Vec.load(word2vec_model_filename)
new_topics = []
useless = []
for topic in topics:
words = topic[-1]
dictionary, matrix = get_words_matrix(words, word2vec)
clusters, centers = cluster(matrix, dictionary, 2, 10)
cohesions = []
for c in clusters.items():
sub_words = c[-1]
label = c[0]
_, sub_matrix = get_words_matrix(sub_words, word2vec)
center = centers[label]
cohesion = utilities.cohesion(sub_matrix, center)
cohesions.append((label, cohesion))
cohesions.sort(key=lambda x: x[-1])
new_topic = list(topic[:-1])
new_topic.append(cohesions[0][1])
new_topic.append(clusters[cohesions[0][0]])
new_topics.append(new_topic)
for c in cohesions[1:]:
u_topic = list(topic[:-1])
u_topic.append(c[0])
u_topic.append(c[1])
u_topic.append(clusters[c[0]])
useless.append(u_topic)
return new_topics, useless
def cluster_analyse_with_cohesion(words, word2vec_model, k=1):
dictionary, matrix = get_words_matrix(words, word2vec_model)
clusters, centers = cluster(matrix, dictionary, 2, 10)
topics = []
t_centers = dict()
depth = 0
cohesions = []
for c in clusters.items():
topics.append(list(c))
label = c[:-1]
t_centers[label] = centers[c[0]]
words = c[-1]
sub_matrix = get_words_matrix(words, word2vec_model)
center = t_centers[label]
cohesion = utilities.cohesion(sub_matrix, center)
cohesions.append((label, cohesion))
if len(c) - 1 > depth:
depth = len(c) - 1
cohesions.sort(key=lambda x: x[-1])
while clusters_score and depth < k:
label = clusters_score[0][0]
del clusters_score[0]
topic = topics[label]
subwords = topic[-1]
dictionary, matrix = get_words_matrix(subwords, word2vec_model)
clusters = cluster(matrix, dictionary, 2, 10)
new_topics = []
for c in clusters:
new_topic = topic[:-1]
new_topic.extend(list(c))
new_topics.append(new_topic)
new_topics.extend(topics[:label])
new_topics.extend(topics[label + 1 :])
new_labels, new_samples_score = silhouette_samples(new_topics, word2vec_model)
clusters_score = silhouette_clusters(new_labels, new_samples_score)
clusters_score.sort(key=lambda x: x[1])
topics = new_topics
for t in topics:
if len(t) - 1 > depth:
depth = len(t) - 1
return topics
