Setting up time steps for forecasting

Now the time series prediction model predicts for a given number of elements (forecast horizon). But for some models, training for long horizons may be ineffective. So it is possible to train a model to predict a small horizon, and then use "out-of-sample" forecasting.


It is required to implement a mechanism that will allow to find a tarde-off between the length of model forecast and the number of steps in out-of-sample forecasting