"""run lda/plsa"""
topic_word, doc_topic = ldaAdd.runLda(lda_m, dic)
user_topic, users, users_pic_num = ldaAdd.userTopic(USER_FILE, points,
doc_topic)
if not (os.path.isfile(ZW_FILE) and os.path.isfile(DZ_FILE)):
plsa_topic_word, plsa_doc_topic = plsaAdd.runPlsa(plsa_m, dic,
CLUS_WORD_ZERO_FILE,
ZW_FILE, DZ_FILE)
else:
plsa_topic_word, plsa_doc_topic = plsaAdd.loadPlsa(ZW_FILE, DZ_FILE,
clus_num, len(dic))
plsa_user_topic, users, users_pic_num = ldaAdd.userTopic(
USER_FILE, points, plsa_doc_topic)
"""trans/clus time, order score"""
sc.estTransOrder(points, users, cluster_centers)
rm_by_clus = points[:, -2] < sc.clus_k
somepoints = []
for x in points[rm_by_clus]:
if x[1] in users:
somepoints.append(x)
somepoints = np.array(somepoints)
print 'somepoints:'
print len(somepoints)
random.seed(100)
randpoints = random.sample(somepoints, 15000)
randpoints = np.array(randpoints)
clus_hr_sort = sc.lmsOfClusHr(users, user_topic, doc_topic, somepoints, [],
midclus.write(str(points[i,-2])+' ')
midclus.close()
#drawGmap.drawLayer(labels2, cluster_centers2, n_clusters_2, loc, 2)
clus_num = n_clusters_2
"""run lda"""
if not os.path.isfile(LDA_FILE):
lda_m = ldaAdd.saveLda(clus_num, dic, points, LDA_FILE, LDA_ZERO_FILE)
else:
lda_m = ldaAdd.readLda(LDA_FILE, LDA_ZERO_FILE)
topic_word, doc_topic = ldaAdd.runLda(lda_m, dic)
user_topic, users, t = ldaAdd.userTopic(USER_FILE, points, doc_topic)
"""trans/clus time, order score"""
sc.estTransOrder(points, users, cluster_centers)
rm_by_clus = points[:,-2] < sc.clus_k
somepoints = []
for x in points[rm_by_clus]:
if x[1] in users:
somepoints.append(x)
somepoints = np.array(somepoints)
print 'somepoints:'
print len(somepoints)
random.seed(100)
randpoints = random.sample(somepoints, 15000)
randpoints = np.array(randpoints)
trainpoints = []