Regularization

[MatteoDellaTorre]

Hello everyone. I have implemented a model with lightweight.

I have noticed that when predicting, varying the input expenditure, the output provided by the model is always very similar.

This does not allow the optimizer to work properly.

In my opinion, the model manages to capture the trend and seasonality well, but gives less weight to the media channels with the associated spend.

Is there a solution to this problem? Could a regularization parameter be introduced by forcing the model to give more importance to the media channels?