def training(self):
filename = path + '\Corpus\{0}_ya.txt'.format(self.target_word)
# ----- loading Wiktionary data -----
self.worddata = WiktionaryData(self.target_word, self.language)
# ---- loading corpus -----
self.corpus = Corpus()
self.corpus.load_corpus(filename, self.language, self.target_word)
count = len(self.corpus.paragraphes)
corpus_paragraphes_count = count
# ----- add definitions and examples to corpus
for d in self.worddata.definitions:
paragraph = Paragraph(count)
d['parnmb'] = count
s = Sentence(d['sent'].fulltext, self.language, self.target_word)
if 'ex' in d:
for exmpl in d['ex']:
s.stemmed_tokens += exmpl.stemmed_tokens
s.tokens += exmpl.tokens
paragraph.add_sentence(s)
paragraph.add_definition(d)
self.corpus.add_paragraph(paragraph)
count += 1
rng = numpy.array(range(corpus_paragraphes_count, len(self.corpus.paragraphes)))
# ---- creating tf-idf matrix -----
self.corpus.create_tf_idf_matrix()
# ----- draw -----
draw2d = Drawing(2, self.corpus.tf_idf_matrix, len(self.worddata.definitions), rng)
draw2d.draw()
draw3d = Drawing(3, self.corpus.tf_idf_matrix, len(self.worddata.definitions), rng)
draw3d.draw()
del(draw2d)
del(draw3d)
# ----- matrix reduction -----
if self.reduction_mode == 'SVD':
self.pca_object = myPCA(0.95, 2)
self.pca_object.count_number_of_principal_components(self.corpus.tf_idf_matrix)
new_X = self.pca_object.reduce_matrix_dimension(self.corpus.tf_idf_matrix)
else:
new_X = self.corpus.tf_idf_matrix
# ----- clustering -----
self.cluster_object = Cluster('K-means', len(self.worddata.definitions), new_X, rng)
self.cluster_object.clustering()
self.cluster_object.make_word_clusters(self.corpus)
print('Trainig is done')
def main():
print("Composite pattern Shapes..")
cir1 = Circle()
cir1.draw("red")
cir2 = Circle()
cir2.draw("blue")
tri1 = Triangle()
tri1.draw("green")
drawing = Drawing()
drawing.add(cir1)
drawing.add(tri1)
drawing.draw("yellow")
drawing.add(cir2)
drawing.draw("orange")
drawing.remove(tri1)
drawing.draw("orange")
