Why did we assign 0 to the threshold in the last method?
I think in this value all nonsignificant features will be included in the model !!!
if that is what will happen, how can I select the optimal value for the threshold.....?
Maybe I misunderstand the variance concept, please correct me if it is