## MULTIPLYING RETWEETS FOLLOWERS
print("multiplying nr. of retweet followers by sentiments.")
sentanalyzer = SentimentAnalyzer()
sentanalyzer.merge_tweets_with_retweets(coin)
#print(coin.tweets)
sentanalyzer.sent_mul_tweet_followers(coin)
sentanalyzer.sent_mul_retweet_followers(coin)
print(len(coin.retweets))
print(coin.retweets.tail())
## GROUPING RETWEETS BY HOUR
print("grouping retweets by hour basis")
sentanalyzer.group_retweet_by_hour(coin)
print(coin.grtdf.head())
print("grouping tweets by hour basis")
sentanalyzer.group_tweet_by_hour(coin)
print(coin.gtdf.tail())
## COIN PRICE
coinprice = CoinPrice()
coinprice.read_and_sort_price(coin)
print(coin.pricehourly.tail())
phase = "prepare2"
coin.save_to_storeage(phase)
'eu_market',
'us_market',
'day',
'max_datetime_x',
'max_datetime_y'
],
inplace=True)
#cointrain.add_log_columns(data,strCols=True)
print(len(data))
print('data.columns')
print(data.columns)
#print(data)
coin.data_to_predict = data
coin.save_to_storeage('train')
#min_max_scaler = preprocessing.MinMaxScaler()
#np_scaled = min_max_scaler.fit_transform(data)
#data = pd.DataFrame(np_scaled)
#coin.save_scaler(min_max_scaler)
#cointrain.add_square_columns(data,strCols=False)
def precision(y_true, y_pred):
threshold = 0.55
mult = 0.5 / threshold
true_positives = K.sum(
K.round(K.clip(y_true * y_pred * mult, 0, 1)))
predicted_positives = K.sum(