GitLab

Created by: Dreamlone

To fill in gaps in the time series, an important hyperparameter is the size of the moving window. To fill in the gaps, use time series forecasting models that need to specify this parameter.

The problem is as follows

In some cases, the window size will not allow us to train the model to fill in the gap, in others, the selected window size will reduce the learning ability of the model. So, there is a need to implement an efficient algorithm for determining the optimal window size and embedded it in a single mechanism for effectively gap-filling.