Previous topic


Next topic


_RollingBasis.__call__(train_period, test_period, start=0, end=None, roll_period=None, eval_period=None, batch_size=64, epochs=1)

Callable method to set target features data, and model.

train_period, test_period : int

Size of respectively training and testing sub-periods.

start : int, optional

Starting observation, default is first observation.

end : int, optional

Ending observation, default is last observation.

roll_period : int, optional

Size of the rolling period, default is the same size of the testing sub-period.

eval_period : int, optional

Size of the evaluating period, default is the same size of the testing sub-period if training sub-period is large enough.

batch_size : int, optional

Size of a training batch, default is 64.

epochs : int, optional

Number of epochs on the same subperiod, default is 1.


The rolling basis model.