def extract(self, train_instance):
pos_sa, pos_sb = train_instance.get_word(type='pos', stopwords=True)
features, infos = utils.sentence_vectorize_features(pos_sa,
pos_sb,
self.idf_weight,
convey='count')
return features, infos
def extract(self, train_instance):
sa, sb = train_instance.get_word(type='lemma',
stopwords=self.stopwords,
lower=True)
features, infos = utils.sentence_vectorize_features(
sa, sb, self.idf_weight)
return features, infos
def extract(self, train_instance):
idf_weight = dict_utils.DictLoader().load_dict('global_idf')
sa, sb = train_instance.get_word(type='lemma',
stopwords=True,
lower=True)
features, infos = utils.sentence_vectorize_features(sa, sb, idf_weight)
return features, infos
def extract(self, train_instance):
dep_sa, dep_sb = train_instance.get_dependency()
features = []
feature, info = utils.sentence_match_features(dep_sa, dep_sb)
features += feature
feature, info = utils.sentence_vectorize_features(dep_sa,
dep_sb,
self.idf_weight,
convey=self.convey)
features += feature
infos = [dep_sa, dep_sb]
return features, infos
def extract(self, train_instance):
sa, sb = train_instance.get_preprocess()
# sa, sb = train_instance.get_word(type='lemma', stopwords=True, lower=True)
la, lb = len(sa), len(sb)
l = min(la, lb)
features = []
feature, info = utils.sentence_sequence_features(sa, sb)
features += feature
feature, info = utils.sentence_match_features(sa, sb)
features += feature
bow = utils.idf_calculator([sa, sb])
feature, info = utils.sentence_vectorize_features(sa,
sb,
bow,
convey='count')
features += feature
infos = [sa, sb]
return features, infos
