if anyone can help I would be very grateful, I use a hybrid method to forecast, the deep learning architecture I use is Bi-LSTM, my data is 2788 data with 2 predictor variables, I use data normalization using a min-max scaler. is it correct?, I have doubts about the nominal distance of the data, namely in this case I used gold and silver prices that were too far away, when making predictions, the forecast results were very far from the original data, did I do something wrong, if anyone knows, please help , Thank You
I have tried modifying the Bi-LSTM model by adding a hidden layer but it doesn't give good results, I want forecasting results that can have forecasting data that fits the original data