Skip to main content

Table 2 LSTM parameters summary

From: Forecasting the carsharing service demand using uni and multivariable models

Model Parameter List
Single Variable { “layers" = [ LSTM_layer(50 Nodes, dropout=0.2), LSTM_layer(50 Nodes, activation=relu), Dense_layer(12 Nodes) ] “optimizer": RMSprop(clipvalue=1.0), “loss": “mae", “epochs": 15, “evaluation_interval":50, “validation_steps": 50, “batch_size": 256 for the three Real-World services and 64 for the Free-floating comparison,}
Multi Variable { “layers" = [ LSTM_layer(80 Nodes, dropout=0.2), LSTM_layer(80 Nodes, activation=relu), Dense_layer(12 Nodes) ] “optimizer": RMSprop(clipvalue=1.0), “loss": “mae", “epochs": 15, “evaluation_interval":50, “validation_steps": 50, “batch_size": 256 for the three Real-World services and 64 for the Free-floating comparison,}