class Model[source]

Model(model_args=None, quantiles=[0.05, 0.5, 0.95], bs=128, epochs=100, lr=0.3, alpha=0.1, silent=True) :: Module

Default Model Wrapper parameters:

• df: dataset used in training dataset.
• model (optional): how to model time series. Default: DefaultModel.
• loss (optional): loss function: Default l1 loss.
• bs (optional): batchsize
• alpha (optional): l2 weight decay
• beta (optional): l1 weight decay
• silent: whether to silence output or not

Usage:

model = Model(train_df)
model.fit(train_df)
y = model.predict(test_df)