def train_docvecs(Sentences):
model = Doc2Vec(min_count=2, window=10, size=50, sample=1e-5, negative=5, workers=7)
model.build_vocab(Sentences.to_array())
for epoch in range(100):
print('epoch: %s' % epoch)
model.train(Sentences.sentences_rand())
model.save(get_file_path('docvecs_CVAT'))
print('Training model complete, saved successful.')
def load_embeddings(arg=None):
if arg == 'zh_tw': # dim = 400
model = gensim.models.Word2Vec.load_word2vec_format(get_file_path('cn_word2vec'), binary=False)
elif arg == 'CVAT': # dim = 50
model = gensim.models.Word2Vec.load(get_file_path('wordvecs_CVAT'))
elif arg == 'IMDb': # dim = 100
model = Doc2Vec.load(get_file_path('test_doc2vec_model'))
elif arg == 'CVAT_docvecs': # dim = 50
model = Doc2Vec.load(get_file_path('docvecs_CVAT'))
elif arg == 'google_news':
model = gensim.models.Word2Vec.load_word2vec_format(get_file_path('google_news'), binary=True)
elif arg == 'vader':
model = gensim.models.Word2Vec.load('./data/vader_wordvecs.w2v')
else:
raise Exception('Wrong Argument.')
print('Load Model Complete.')
return model
def load_embeddings(arg=None):
if arg == 'zh_tw': # dim = 400
model = gensim.models.Word2Vec.load_word2vec_format(
get_file_path('cn_word2vec'), binary=False)
elif arg == 'CVAT': # dim = 50
model = gensim.models.Word2Vec.load(get_file_path('wordvecs_CVAT'))
elif arg == 'IMDb': # dim = 100
model = Doc2Vec.load(get_file_path('test_doc2vec_model'))
elif arg == 'CVAT_docvecs': # dim = 50
model = Doc2Vec.load(get_file_path('docvecs_CVAT'))
elif arg == 'google_news':
model = gensim.models.Word2Vec.load_word2vec_format(
get_file_path('google_news'), binary=True)
elif arg == 'vader':
model = gensim.models.Word2Vec.load('./data/vader_wordvecs.w2v')
else:
raise Exception('Wrong Argument.')
print('Load Model Complete.')
return model
def log_performance(MSE, MAE, Pearson_r, R2, Spearman_r, sqrt_MSE):
# create a file handler
handler = logging.FileHandler(get_file_path('log'))
handler.setLevel(logging.INFO)
# create a logging format
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
handler.setFormatter(formatter)
# add the handlers to the logger
logger.addHandler(handler)
logger.info('MSE: %s, MAE: %s, Pearson_r: %s, R2: %s, Spearman_r: %s, sqrt_MSE: %s', MSE, MAE, Pearson_r, R2, Spearman_r, sqrt_MSE)
logger.removeHandler(handler) # remove the Handler after you finish your job
def log_state(msg):
# create a file handler
handler = logging.FileHandler(get_file_path('log'))
handler.setLevel(logging.INFO)
# create a logging format
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
handler.setFormatter(formatter)
# add the handlers to the logger
logger.addHandler(handler)
logger.info(msg)
logger.removeHandler(handler)
def train_wordvecs(Sentence, save_path=None):
model = Word2Vec(size=50, min_count=1)
model.build_vocab(Sentence.toarray())
for epoch in range(10):
print('epoch: %s' % epoch)
model.train(Sentence.rand())
if save_path is None:
model.save(get_file_path('wordvecs_CVAT'))
else:
model.save(save_path)
print('Training model complete, saved successful.')
def train_wordvecs(Sentence, save_path=None):
model = Word2Vec(size=50, min_count=1)
model.build_vocab(Sentence.toarray())
for epoch in range(10):
print('epoch: %s' % epoch)
model.train(Sentence.rand())
if save_path is None:
model.save(get_file_path('wordvecs_CVAT'))
else:
model.save(save_path)
print('Training model complete, saved successful.')
def log_state(msg):
# create a file handler
handler = logging.FileHandler(get_file_path('log'))
handler.setLevel(logging.INFO)
# create a logging format
formatter = logging.Formatter(
'%(asctime)s - %(name)s - %(levelname)s - %(message)s')
handler.setFormatter(formatter)
# add the handlers to the logger
logger.addHandler(handler)
logger.info(msg)
logger.removeHandler(handler)
def train_docvecs(Sentences):
model = Doc2Vec(min_count=2,
window=10,
size=50,
sample=1e-5,
negative=5,
workers=7)
model.build_vocab(Sentences.to_array())
for epoch in range(100):
print('epoch: %s' % epoch)
model.train(Sentences.sentences_rand())
model.save(get_file_path('docvecs_CVAT'))
print('Training model complete, saved successful.')
def log_performance(MSE, MAE, Pearson_r, R2, Spearman_r, sqrt_MSE):
# create a file handler
handler = logging.FileHandler(get_file_path('log'))
handler.setLevel(logging.INFO)
# create a logging format
formatter = logging.Formatter(
'%(asctime)s - %(name)s - %(levelname)s - %(message)s')
handler.setFormatter(formatter)
# add the handlers to the logger
logger.addHandler(handler)
logger.info(
'MSE: %s, MAE: %s, Pearson_r: %s, R2: %s, Spearman_r: %s, sqrt_MSE: %s',
MSE, MAE, Pearson_r, R2, Spearman_r, sqrt_MSE)
logger.removeHandler(
handler) # remove the Handler after you finish your job
import numpy as np
def get_vocab(corpus):
vocab = defaultdict(float)
for sent in corpus:
for word in sent:
vocab[word] += 1
print(len(vocab))
return vocab
########################################## config ########################################
vec_dim = 400
##########################################################################################
corpus = load_corpus(get_file_path('cn_corpus'))
# print(corpus[:2])
# vocab = get_vocab(corpus)
# dump_picle(vocab, get_file_path('CVAT_Vocab'))
# print('OK')
vocab = load_pickle(get_file_path('CVAT_Vocab'))
# for i in vocab:
# print(i)
# print(len(vocab))
# W, word_idx_map = build_embedding_matrix(load_embeddings('zh_tw'), vocab, k=400)
# dump_picle(word_idx_map, get_file_path('word_idx_map_CVAT'))
# print('dump word_idx_map successful')
# dump_picle(W, '/home/hs/Data/embedding_matrix_CVAT.p')
# print('OK')
from positive_negative_split import get_pos_neg_va
def cv(data, target, multivariant=False):
X_train, X_test, Y_train, Y_test = cross_validation.train_test_split(
data, target, test_size=0.2, random_state=0)
if multivariant is False:
linear_regression(X_train, X_test, Y_train, Y_test, plot=False)
else:
linear_regression_multivariant(X_train,
X_test,
Y_train,
Y_test,
cost_fun='Ridge_Regression')
normalize = True
corpus = load_corpus(get_file_path('cn_corpus'))
lexicon = load_lexicon(get_file_path('normalized_onezero_lexicon'))
mark = load_mark(get_file_path('normalized_onezero_mark'))
# # the following could use to check the same words in corpus and lexicon
# from visualization import show_common_term
# show_common_term(corpus, lexicon)
# exit()
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_sqr(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = nonlinear_max_fusion(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_tf(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_tfidf(corpus, lexicon, mark)
valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_geo(
corpus, lexicon, mark)
# get vocab and save to pickle
vocab = get_vocab(file_dir)
dump_picle(vocab, './data/tmp/vocab.p')
print('OK')
a = load_pickle('./data/tmp/vocab.p')
for i in a:
print(i)
print(len(a))
exit()
# end
# make word index map
vocab = load_pickle('./data/tmp/vocab.p')
W, word_idx_map = build_embedding_matrix(load_embeddings('google_news'), vocab, k=300)
dump_picle(word_idx_map, get_file_path('word_idx_map'))
print('dump word_idx_map successful')
dump_picle(W, '/home/hs/Data/embedding_matrix.p')
print('OK')
exit()
# make word index map end
word_idx_map = load_pickle(get_file_path('word_idx_map'))
print(len(word_idx_map))
for i in word_idx_map:
print(i)
exit()
word_idx_map = load_pickle(get_file_path('word_idx_map'))
data, pos_length, neg_length = prepare_data(file_dir, word_idx_map)
dump_picle([data, pos_length, neg_length], get_file_path('imdb_processed_data'))
if i % 10 == 0:
logger.info("evaluate for text : %i/%i..." % (i, num))
evaluate(valence_true, valence_pred, 'valence')
evaluate(arousal_true, arousal_pred, 'arousal')
if __name__ == '__main__':
program = os.path.basename(sys.argv[0])
logger = logging.getLogger(program)
logging.basicConfig(format='%(asctime)s: %(levelname)s: %(message)s')
logging.root.setLevel(level=logging.INFO)
logger.info(r"running %s" % ''.join(sys.argv))
corpus_name = get_file_path('cn_corpus')
logger.info(r"loading corpus from : " + corpus_name)
lexicon_name = get_file_path('lexicon')
logger.info(r"loading lexicon form : " + lexicon_name)
expand_name = get_file_path('neural_cand')
logger.info(r"loading expand_word from : " + expand_name)
mark_name = get_file_path('mark')
logger.info(r"loading mark from : " + mark_name)
corpus = load_corpus(corpus_name)
lexicon = load_lexicon(lexicon_name)
mark = load_mark(mark_name)
# log_state('use extend lexicon')
if i % 10 == 0:
logger.info("evaluate for text : %i/%i..." % (i, num))
evaluate(valence_true, valence_pred, 'valence')
evaluate(arousal_true, arousal_pred, 'arousal')
if __name__ == '__main__':
program = os.path.basename(sys.argv[0])
logger = logging.getLogger(program)
logging.basicConfig(format='%(asctime)s: %(levelname)s: %(message)s')
logging.root.setLevel(level=logging.INFO)
logger.info(r"running %s" % ''.join(sys.argv))
corpus_name = get_file_path('cn_corpus')
logger.info(r"loading corpus from : " + corpus_name)
lexicon_name = get_file_path('lexicon')
logger.info(r"loading lexicon form : " + lexicon_name)
expand_name = get_file_path('neural_cand')
logger.info(r"loading expand_word from : " + expand_name)
mark_name = get_file_path('mark')
logger.info(r"loading mark from : " + mark_name)
corpus = load_corpus(corpus_name)
lexicon = load_lexicon(lexicon_name)
mark = load_mark(mark_name)
# log_state('use extend lexicon')
for word in clean_str(sent).split():
vocab[word] += 1
print(len(vocab))
return vocab
def process(corpus):
return [clean_str(sent) for sent in corpus]
vec_dim = 300
############################################## all ###############################################
corpus, ratings = load_vader(['tweets', 'movie_reviews', 'product_reviews', 'news_articles'])
# corpus, ratings = load_vader(['movie_reviews'])
# # name: tweets, movie_reviews, product_reviews, news_articles
lexicon_name = get_file_path('anew')
words, valences, _ = load_anew(lexicon_name)
corpus, ratings = screen_data(corpus, ratings, words)
ratings = np.array(ratings) + np.ones(len(ratings), dtype=float) * 5
print(len(corpus), len(corpus))
vocab = get_vocab(corpus)
# dump_picle(vocab, './data/corpus/vader/vocab_moview_reviews.p')
# vocab = load_pickle('./data/corpus/vader/vocab_moview_reviews.p')
print(len(vocab))
W, word_idx_map = build_embedding_matrix(load_embeddings('google_news'), vocab, k=300)
# dump_picle(word_idx_map, './data/corpus/vader/word_idx_map_movie_reviews.p')
# print('dump word_idx_map successful')
dump_picle(W, './data/corpus/vader/embedding_matrix_all.p')
print('dump embedding matrix file OK')
c
from file_name import get_file_path
from load_data import load_corpus, load_lexicon, load_mark
from load_data import load_embeddings
from word2vec_fn import buill_word_vector
from word2vec_fn import gold_valence_arousal
import numpy as np
from sklearn import cross_validation
from cross_validation import cv
from word2vec_fn import build_doc_vector
# '''
model = load_embeddings('CVAT_docvecs')
print(model.docvecs[1])
print(model.docvecs['SENT_23'])
print(len(model.vocab.keys()))
corpus = load_corpus(get_file_path('cn_corpus'))
mark = load_mark(get_file_path('mark'))
vecs = build_doc_vector(corpus, model)
valence, arousal = gold_valence_arousal(corpus, mark)
cv(vecs, valence, multivariant=True)
cv(vecs, arousal, multivariant=True)
# '''
# from save_data import dump_picle
# dump_picle(model.key(), get_file_path('words_in_wordvec'))
# print('ok')
#
# # print(model.most_similar(positive=['woman', 'king'], negative=['man'], topn=1))
# # print(model.doesnt_match("breakfast cereal dinner lunch".split()))
# # print(model.similarity('woman', 'man'))
# # print(model.most_similar_cosmul(positive=['baghdad', 'england'], negative=['london'], topn=10))
for item_no, line in enumerate(fin):
self.sentences.append(
LabeledSentence(words=utils.to_unicode(line).split(), tags=[prefix + '_%s' % str(item_no)]))
return self.sentences
def sentences_rand(self):
# out = numpy.random.permutation(self.sentences)
# return out
random.shuffle(self.sentences)
return self.sentences
sources = {'test-neg.txt': 'TEST_NEG', 'test-pos.txt': 'TEST_POS', 'train-neg.txt': 'TRAIN_NEG',
'train-pos.txt': 'TRAIN_POS', 'train-unsup.txt': 'TRAIN_UNS'}
##########################################################
dir = get_file_path('test_doc2vec')
keys = list(sources.keys())
for old_key in keys:
sources[dir + '/' + old_key] = sources.pop(old_key)
##############################################################
############################# model training #################
'''
sentences = LabeledLineSentence(sources)
model = Doc2Vec(min_count=1, window=10, size=100, sample=1e-4, negative=5, workers=7)
model.build_vocab(sentences.to_array())
for epoch in range(10):
print('epoch: %s' % epoch)
model.train(sentences.sentences_rand())
print(model.most_similar('good'))
# print(model['TRAIN_NEG_0'])
def cv(data, target, multivariant=False):
X_train, X_test, Y_train, Y_test = cross_validation.train_test_split(data, target, test_size=0.1, random_state=10)
if multivariant is False:
linear_regression(X_train, X_test, Y_train, Y_test, plot=False)
else:
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun="ordinary_least_squares")
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun="Ridge_Regression")
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun="Bayesian_Regression")
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun="SVR")
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun="KNN_Reg")
if __name__ == "__main__":
normalize = True
corpus = load_corpus(get_file_path("cn_corpus"))
# lexicon = load_lexicon(get_file_path('lexicon'))
mark = load_mark(get_file_path("mark"))
lexicon = combine_lexicon(get_file_path("lexicon"), get_file_path("neural_cand"))
# # the following could use to check the same words in corpus and lexicon
# from visualization import show_common_term
# show_common_term(corpus, lexicon)
# exit()
valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion(corpus, lexicon, mark)
print("start.....")
cv(valence_mean, valence_true, multivariant=False)
cv(arousal_mean, arousal_true, multivariant=False)
print("OK")
def sentences_rand(self):
# out = numpy.random.permutation(self.sentences)
# return out
random.shuffle(self.sentences)
return self.sentences
sources = {
'test-neg.txt': 'TEST_NEG',
'test-pos.txt': 'TEST_POS',
'train-neg.txt': 'TRAIN_NEG',
'train-pos.txt': 'TRAIN_POS',
'train-unsup.txt': 'TRAIN_UNS'
}
##########################################################
dir = get_file_path('test_doc2vec')
keys = list(sources.keys())
for old_key in keys:
sources[dir + '/' + old_key] = sources.pop(old_key)
##############################################################
############################# model training #################
'''
sentences = LabeledLineSentence(sources)
model = Doc2Vec(min_count=1, window=10, size=100, sample=1e-4, negative=5, workers=7)
model.build_vocab(sentences.to_array())
for epoch in range(10):
print('epoch: %s' % epoch)
model.train(sentences.sentences_rand())
print(model.most_similar('good'))
# print(model['TRAIN_NEG_0'])
__author__ = "NLP-PC"
# coding: utf-8
import goslate
from load_data import load_lexicon
from file_name import get_file_path
from load_data import load_anew
import os
os.chdir("..")
print(load_lexicon(get_file_path("lexicon")))
words, valence, arousal = load_anew(get_file_path("anew"))
gs = goslate.Goslate()
for tw_text in gs.translate(words, "zh-tw"):
print(tw_text)
print(gs.translate("Hi", "zh-TW"))
# You could get all supported language list through get_languages
languages = gs.get_languages()
print(languages["zh-TW"])
def get_vocab(corpus):
vocab = defaultdict(float)
for sent in corpus:
for word in sent:
vocab[word] += 1
print(len(vocab))
return vocab
# 注意： 这个文件就是CVAT构造cnn输入数据的代码
########################################## config ########################################
vec_dim = 400
##########################################################################################
corpus = load_corpus(get_file_path('cn_corpus'))
print(corpus[:2])
vocab = get_vocab(corpus)
dump_picle(vocab, get_file_path('CVAT_Vocab'))
print('Dump CVAT vocab OK')
# vocab = load_pickle(get_file_path('CVAT_Vocab'))
for i in vocab:
print(i)
print(len(vocab))
W, word_idx_map = build_embedding_matrix(load_embeddings('zh_tw'),
vocab,
k=400)
dump_picle(word_idx_map, get_file_path('word_idx_map_CVAT'))
print('dump word_idx_map successful')
dump_picle(W, './data/tmp/embedding_matrix_CVAT.p')
raise Exception('Wrong Argument.')
print('Load Model Complete.')
return model
def load_vader(name):
def load_text(filename):
with open(filename, 'r', encoding='utf-8') as csvfile:
reader = csv.reader(csvfile, delimiter='\t')
texts, ratings = [], []
for line in reader:
texts.append(line[2])
ratings.append(float(line[1]))
return texts, ratings
texts, ratings = [], []
for filename in name:
text, rating = load_text('./data/corpus/vader/' + filename + '.txt')
texts.extend(text)
ratings.extend(rating)
return texts, ratings
if __name__ == '__main__':
from file_name import get_file_path
words = load_corpus(get_file_path('cn_corpus'))
print(words[719])
# for i, w in enumerate(words):
# print(i,w)
cost_fun='Ridge_Regression')
if __name__ == '__main__':
from load_data import load_vader
normalize = True
corpus, ratings = load_vader(['movie_reviews'])
corpus = process(corpus)
# lexicon = load_lexicon(get_file_path('lexicon'))
from load_data import load_anew
from file_name import get_file_path
import numpy as np
words, valences, _ = load_anew(get_file_path('anew'))
mark = np.array(ratings) + np.ones(len(ratings), dtype=float) * 5
lexicon = dict()
for i, word in enumerate(words):
lexicon[word] = valences[i]
# # the following could use to check the same words in corpus and lexicon
# from visualization import show_common_term
# show_common_term(corpus, lexicon)
# exit()
# valence_mean, valence_true = linear_fusion(corpus, lexicon, mark)
# print('start.....')
# cv(valence_mean, valence_true, multivariant=False)
# print('OK')
import numpy as np
def statistic(texts):
avg_length, vocab = 0, set()
length_list = []
for text in texts:
if type(text) is not list:
text = text.split()
length_list.append(len(text))
vocab = vocab.union(set(text))
# if len(text)>200:
# print(text)
avg_length = np.average(np.array(length_list))
return avg_length, len(vocab)
if __name__ == '__main__':
# (['tweets', 'movie_reviews', 'product_reviews', 'news_articles'])
tweets, _ = load_vader(['tweets'])
movie, _ = load_vader(['movie_reviews'])
amazon, _ = load_vader(['product_reviews'])
NYT, _ = load_vader(['news_articles'])
cvat = load_corpus(get_file_path('cn_corpus'))
print(statistic(tweets))
print(statistic(movie))
print(statistic(amazon))
print(statistic(NYT))
print(statistic(cvat))
return result
def scaling_onezero(num_list):
# Note: the type of the parameter is np.array
# Function: To normalize data
result = []
for num in num_list:
result.append(num / 9)
return result
if __name__ == '__main__':
from load_data import load_lexicon
from load_data import load_mark
from file_name import get_file_path
from save_data import save_csv
lexicon = load_lexicon(get_file_path('lexicon'))
mark = load_mark(get_file_path('mark'))
lexicon = np.array(lexicon)
mark = np.array(mark)
#####################################
lexicon[:, 1] = scaling_onezero(np.array(lexicon[:, 1], dtype=float))
lexicon[:, 2] = scaling_onezero(np.array(lexicon[:, 2], dtype=float))
mark[:, 1] = scaling_onezero(np.array(mark[:, 1], dtype=float))
mark[:, 2] = scaling_onezero(np.array(mark[:, 2], dtype=float))
######################################
save_csv(lexicon, get_file_path('normalized_onezero_lexicon'))
save_csv(mark, get_file_path('normalized_onezero_mark'))
vocab[word] += 1
print(len(vocab))
return vocab
def process(corpus):
return [clean_str(sent) for sent in corpus]
vec_dim = 300
############################################## all ###############################################
corpus, ratings = load_vader(
['tweets', 'movie_reviews', 'product_reviews', 'news_articles'])
# corpus, ratings = load_vader(['movie_reviews'])
# # name: tweets, movie_reviews, product_reviews, news_articles
lexicon_name = get_file_path('anew')
words, valences, _ = load_anew(lexicon_name)
corpus, ratings = screen_data(corpus, ratings, words)
ratings = np.array(ratings) + np.ones(len(ratings), dtype=float) * 5
print(len(corpus), len(corpus))
vocab = get_vocab(corpus)
# dump_picle(vocab, './data/corpus/vader/vocab_moview_reviews.p')
# vocab = load_pickle('./data/corpus/vader/vocab_moview_reviews.p')
print(len(vocab))
W, word_idx_map = build_embedding_matrix(load_embeddings('google_news'),
vocab,
k=300)
# dump_picle(word_idx_map, './data/corpus/vader/word_idx_map_movie_reviews.p')
# print('dump word_idx_map successful')
dump_picle(W, './data/corpus/vader/embedding_matrix_all.p')
from file_name import get_file_path
from regression import linear_regression, linear_regression_multivariant
from positive_negative_split import get_pos_neg_va
def cv(data, target, multivariant=False):
X_train, X_test, Y_train, Y_test = cross_validation.train_test_split(data, target, test_size=0.2,
random_state=0)
if multivariant is False:
linear_regression(X_train, X_test, Y_train, Y_test, plot=False)
else:
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='Ridge_Regression')
normalize = True
corpus = load_corpus(get_file_path('cn_corpus'))
lexicon = load_lexicon(get_file_path('normalized_onezero_lexicon'))
mark = load_mark(get_file_path('normalized_onezero_mark'))
# # the following could use to check the same words in corpus and lexicon
# from visualization import show_common_term
# show_common_term(corpus, lexicon)
# exit()
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_sqr(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = nonlinear_max_fusion(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_tf(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_tfidf(corpus, lexicon, mark)
valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_geo(corpus, lexicon, mark)
# valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion_geo_tf(corpus, lexicon, mark)
def cv(data, target, multivariant=False):
X_train, X_test, Y_train, Y_test = cross_validation.train_test_split(data, target, test_size=0.1, random_state=10)
if multivariant is False:
linear_regression(X_train, X_test, Y_train, Y_test, plot=False)
else:
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='ordinary_least_squares')
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='Ridge_Regression')
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='Bayesian_Regression')
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='SVR')
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='KNN_Reg')
if __name__ == '__main__':
normalize = True
corpus = load_corpus(get_file_path('cn_corpus'))
# lexicon = load_lexicon(get_file_path('lexicon'))
mark = load_mark(get_file_path('mark'))
lexicon = combine_lexicon(get_file_path('lexicon'), get_file_path('neural_cand'))
# # the following could use to check the same words in corpus and lexicon
# from visualization import show_common_term
# show_common_term(corpus, lexicon)
# exit()
valence_mean, valence_true, arousal_mean, arousal_true = linear_fusion(corpus, lexicon, mark)
print('start.....')
cv(valence_mean, valence_true, multivariant=False)
cv(arousal_mean, arousal_true, multivariant=False)
print('OK')
raise Exception('Wrong Argument.')
print('Load Model Complete.')
return model
def load_vader(name):
def load_text(filename):
with open(filename, 'r', encoding='utf-8') as csvfile:
reader = csv.reader(csvfile, delimiter='\t')
texts, ratings = [], []
for line in reader:
texts.append(line[2])
ratings.append(float(line[1]))
return texts, ratings
texts, ratings = [], []
for filename in name:
text, rating = load_text('./data/corpus/vader/' + filename + '.txt')
texts.extend(text)
ratings.extend(rating)
return texts, ratings
if __name__ == '__main__':
from file_name import get_file_path
words = load_corpus(get_file_path('cn_corpus'))
print(words[719])
# for i, w in enumerate(words):
# print(i,w)
linear_regression_multivariant(X_train, X_test, Y_train, Y_test, cost_fun='Ridge_Regression')
if __name__ == '__main__':
from load_data import load_vader
normalize = True
corpus, ratings = load_vader(['movie_reviews'])
corpus = process(corpus)
# lexicon = load_lexicon(get_file_path('lexicon'))
from load_data import load_anew
from file_name import get_file_path
import numpy as np
words, valences, _ = load_anew(get_file_path('anew'))
mark = np.array(ratings) + np.ones(len(ratings), dtype=float) * 5
lexicon = dict()
for i, word in enumerate(words):
lexicon[word] = valences[i]
# # the following could use to check the same words in corpus and lexicon
# from visualization import show_common_term
# show_common_term(corpus, lexicon)
# exit()
# valence_mean, valence_true = linear_fusion(corpus, lexicon, mark)
# print('start.....')
# cv(valence_mean, valence_true, multivariant=False)
# print('OK')
vocab = get_vocab(file_dir)
dump_picle(vocab, './data/tmp/vocab.p')
print('OK')
a = load_pickle('./data/tmp/vocab.p')
for i in a:
print(i)
print(len(a))
exit()
# end
# make word index map
vocab = load_pickle('./data/tmp/vocab.p')
W, word_idx_map = build_embedding_matrix(load_embeddings('google_news'),
vocab,
k=300)
dump_picle(word_idx_map, get_file_path('word_idx_map'))
print('dump word_idx_map successful')
dump_picle(W, '/home/hs/Data/embedding_matrix.p')
print('OK')
exit()
# make word index map end
word_idx_map = load_pickle(get_file_path('word_idx_map'))
print(len(word_idx_map))
for i in word_idx_map:
print(i)
exit()
word_idx_map = load_pickle(get_file_path('word_idx_map'))
data, pos_length, neg_length = prepare_data(file_dir, word_idx_map)
dump_picle([data, pos_length, neg_length],
import numpy as np
def get_vocab(corpus):
vocab = defaultdict(float)
for sent in corpus:
for word in sent:
vocab[word] += 1
print(len(vocab))
return vocab
# 注意： 这个文件就是CVAT构造cnn输入数据的代码
########################################## config ########################################
vec_dim = 400
##########################################################################################
corpus = load_corpus(get_file_path('cn_corpus'))
print(corpus[:2])
vocab = get_vocab(corpus)
dump_picle(vocab, get_file_path('CVAT_Vocab'))
print('Dump CVAT vocab OK')
# vocab = load_pickle(get_file_path('CVAT_Vocab'))
for i in vocab:
print(i)
print(len(vocab))
W, word_idx_map = build_embedding_matrix(load_embeddings('zh_tw'), vocab, k=400)
dump_picle(word_idx_map, get_file_path('word_idx_map_CVAT'))
print('dump word_idx_map successful')
dump_picle(W, './data/tmp/embedding_matrix_CVAT.p')
print('OK')
__author__ = 'NLP-PC'
# coding: utf-8
import goslate
from load_data import load_lexicon
from file_name import get_file_path
from load_data import load_anew
import os
os.chdir('..')
print(load_lexicon(get_file_path('lexicon')))
words, valence, arousal = load_anew(get_file_path('anew'))
gs = goslate.Goslate()
for tw_text in gs.translate(words, 'zh-tw'):
print(tw_text)
print(gs.translate('Hi', 'zh-TW'))
# You could get all supported language list through get_languages
languages = gs.get_languages()
print(languages['zh-TW'])
