n = config.list_alltf_result # 各文書における総単語のtf値が入ったリストを代入
"実行したら保存したファイル名を変えること"
fw = open("/home/matsui-pc/matsui/experiment/plsa_n.csv" ,"wa") # 始めは新しく作りその後は追加で書き込んでいく
csvWriter = csv.writer(fw) # csvファイルに書き込むための準備
for i in range(len(n)):
csvWriter.writerow(n[i]) # ファイル名と特徴量をcsvファイルへ書き込む
fw.close()
# print len(n[0])
## 確認用
# for i in range(len(n)):
# print n[i]
# print len(n[0])
p = plsa.plsa(n) # plsa.py の関数plsaに投げる
print "PLSA計算中"
p.train() # EMステップを繰り返す
print "\n"
# print "*********** 最終的な出力 ************"
print "P(z) = ",
print p.pz # P(z)
print "P(d|z) = ",
print p.pz_d # P(d|z)
print "P(w|z) = ",
print p.pz_w # P(w|z)
print "P(z|d,w)",
print p.pdw_z # P(z|d,w)
# iterate over the files in the directory.
document_paths = ['./Files/']
documents = []
for document_path in document_paths:
for document_file in glob.glob(os.path.join(document_path, '*.txt')):
words, lines = common_utils.split(stop_words_set,
document_file) # tokenize
documents.append(words)
vocabulary = common_utils.build_vocabulary(documents)
number_of_topics = 3
max_iterations = 1
topic_word_prob, document_topic_prob = plsa.plsa(number_of_topics,
max_iterations, documents)
common_utils.print_topic_word_distribution(topic_word_prob, vocabulary,
number_of_topics, 3,
"./topic-word.txt")
common_utils.print_document_topic_distribution(document_topic_prob, documents,
number_of_topics, 3,
"./document-topic.txt")
path_wordsim = './wordsim353_sim_rel/wordsim_similarity_goldstandard.txt'
data_cos = []
data_scalar = []
plsa_matrix = pd.DataFrame(data=topic_word_prob, columns=vocabulary)
consistent_wordsim = common_utils.read_consistent_wordsim(
for i in range(nD):
for word in word_freq[i]:
Ndw[i][wordID[word]] = word_freq[i][word]
print Ndw
#pprint(total_freq)
#print Ndw
nZ = 3
noise = +np.random.rand(nD,nW)
Pd_z, Pw_z,Pz_d,Pz_w = plsa.plsa(Ndw+noise,nZ,100)
Y = np.concatenate((Pz_d.T,Pz_w.T))
Y = Y[:,:-1]
#for i in range(len(Y)):
# Y[i] = Y[i][:2]
#Y = Y[:2]
#print np.shape(Y)
#Y = tsne.tsne(Ndw.T,2,nD)
