Actual prediction using inverse_transform for time series data

Question

I would like to apply normalization with min max normalization or Zscore normalization on a single hidden layer of MLP following the example of machine learning mastery for tim