if __name__ == '__main__':
## read bull and bear
bear_fp = "/Users/fredzheng/Documents/stocktwits/sentiment/Bearish"
bull_fp = "/Users/fredzheng/Documents/stocktwits/sentiment/Bullish"
data_fp = "/Users/fredzheng/Documents/stocktwits/sentiment/data_object.pkl"
w2v_fp = "/Users/fredzheng/Documents/stocktwits/sentiment/word2vec"
load_existing_object = True
if load_existing_object:
con = open(data_fp, "rb")
data = pickle.load(con)
con.close()
else:
data = Data()
data.loads(bull_fp, bear_fp, max_n=50000)
data.clean()
data.cut_train_and_test(balance=True)
# data.save(data_fp)
## prep data
tweets_train = data.train
trainX = [tweet.words for tweet in tweets_train]
trainY = [tweet.label for tweet in tweets_train]
train = list(zip(trainX, trainY))
tweets_test = data.test
testX = [tweet.words for tweet in tweets_test]
testY = [tweet.label for tweet in tweets_test]
test = list(zip(testX, testY))
## naive bayes
