# Import data data_source = 'git' market = 'AEX' stocks = get_data(data_source, market) # ONLY FOR NOW, SHOULD BE CHANGED!! df = stocks['PHIA'] # Preprocessing data split_datapoint = 5000 smoothing_window_size = 1000 pp_data = PreProc(df) pp_data.splitdata(split_datapoint) pp_data.normalize_smooth(smoothing_window_size, EMA=0.0, gamma=0.1) # ============================================================================= # Define and apply LSTM # ============================================================================= # Define hyperparameters D = 1 # Dimensionality of the data. Since our data is 1-D this would be 1 num_unrollings = 50 # Number of time steps you look into the future. batch_size = 500 # Number of samples in a batch num_nodes = [200, 200, 150] # Number of hidden nodes in each layer of the deep LSTM stack we're using n_layers = len(num_nodes) # number of layers dropout = 0.2 # Dropout amount # Run LSTM x_axis_seq, predictions_over_time = LSTM(pp_data, D, num_unrollings, batch_size, num_nodes, n_layers, dropout)