dataframe['Date'] = pd.to_datetime(dataframe['Date'])
df_relevant = DLmodels.relevant_stocks(stock,main_df, cor)
df_relevant.index = pd.to_datetime(df_relevant.index)
dataframe = dataframe.merge(df_relevant, how='inner', on='Date')
dataframe = dataframe.merge(datagrouped, how='inner', on='Date')
dataset = DLmodels.clean_dataset(dataframe.drop(['Date'], axis=1)).values
scaler = StandardScaler()
dataset = scaler.fit_transform(dataset)
scaler_filename = 'scalers/Standard_daily_' + stock + '_' + interval + '.save'
pickle.dump(scaler, open(scaler_filename, 'wb'))
scaler = MinMaxScaler(feature_range=[0,1])
dataset = scaler.fit_transform(dataset)
scaler_filename = 'scalers/MinMax_daily_' + stock + '_' + interval + '.save'
pickle.dump(scaler, open(scaler_filename, 'wb'))
X, y = DLmodels.split_sequences(dataset[:-samples_test], n_steps_in, n_steps_out)
n_features = X.shape[2]
y = y[:,:,-1:]
DLmodels.model_LSTM('daily_'+stock, X, y, interval, n_steps_in, n_steps_out, epochs, save, update, verbose)
DLmodels.model_BidirectionalLSTM('daily_'+stock, X, y, interval, n_steps_in, n_steps_out, epochs, save, update, verbose)
DLmodels.model_convLSTM1D('daily_'+stock, X, y, interval, n_steps_in, n_steps_out, epochs, save, update, verbose)
DLmodels.model_ConvLSTM2D('daily_'+stock, X, y, interval, n_steps_in, n_steps_out, epochs, save, update, verbose)
df_target['Date'] = pd.to_datetime(df_target['Date'])
dataframe = df_target
dataframe = dataframe.merge(df_trend_stock_en_us, how="inner", on="Date")
dataframe = dataframe.merge(df_trend_stock_pt_br, how="inner", on="Date")
dataset = dataframe.drop(['Date'], axis=1).dropna().ffill().values
scaler = StandardScaler()
dataset = scaler.fit_transform(dataset)
scaler_filename = 'scalers/complete_' + stock + '_complete_' + interval + '.save'
scaler = pickle.load(open(scaler_filename, 'rb'))
dataset = scaler.fit_transform(dataset)
X, y = DLmodels.split_sequences(dataset, n_steps_in, n_steps_out)
X = X[:, :, :]
n_features = X.shape[2]
y = y[:, :, -1:]
for ML_tech in ML_Techniques:
if ML_tech == 'lr':
(predictions,
accuracy) = DLmodels.stock_forecasting(dataframe, n_steps_out)
if ML_tech == 'lrLog':
(predictions,
accuracy) = DLmodels.stock_forecastingLog(dataframe, n_steps_out)
if ML_tech == 'conv':
predictions = DLmodels.predict_conv1D(