def clustering_word2vec(data_features):
n_start = int(len(data_features)/2)
n_stop = len(data_features)-5
step = 5
n_cls_range = range(n_start, n_stop, step)
n_cls = silhoutte.number_clusters(data_features, n_cls_range)
spectral = KMeans(n_clusters=n_cls).fit(data_features)
label = spectral.fit_predict(data_features)
return label
def clustering_word2vec(data_features):
n_start = int(len(data_features) / 2)
n_stop = len(data_features) - 5
step = 5
n_cls_range = range(n_start, n_stop, step)
n_cls = silhoutte.number_clusters(data_features, n_cls_range)
spectral = KMeans(n_clusters=n_cls).fit(data_features)
label = spectral.fit_predict(data_features)
return label
def clustering(list_perproblem, lang):
vectorizer = TfidfVectorizer(analyzer="char", tokenizer=None, preprocessor=None, stop_words=None,
max_features=5000, min_df=2, ngram_range=(3, 8))
clean_text = []
for j in xrange(0, len(list_perproblem)):
clean_text.append(list_perproblem[j])
data_features = vectorizer.fit_transform(clean_text)
data_features = data_features.toarray()
n_start = int(len(data_features)/2)
n_stop = len(data_features)-5
step = 5
n_cls_range = range(n_start, n_stop, step)
n_cls = silhoutte.number_clusters(data_features, n_cls_range)
spectral = KMeans(n_clusters=n_cls).fit(data_features)
label = spectral.fit_predict(data_features)
return data_features, label
def clustering(list_perproblem, lang):
vectorizer = TfidfVectorizer(analyzer="char",
tokenizer=None,
preprocessor=None,
stop_words=None,
max_features=5000,
min_df=2,
ngram_range=(3, 8))
clean_text = []
for j in xrange(0, len(list_perproblem)):
clean_text.append(list_perproblem[j])
data_features = vectorizer.fit_transform(clean_text)
data_features = data_features.toarray()
n_start = int(len(data_features) / 2)
n_stop = len(data_features) - 5
step = 5
n_cls_range = range(n_start, n_stop, step)
n_cls = silhoutte.number_clusters(data_features, n_cls_range)
spectral = KMeans(n_clusters=n_cls).fit(data_features)
label = spectral.fit_predict(data_features)
return data_features, label