Hi thanks for the comment. The lag 1 day refers to the way the training data are constructed. E.g. by looking at the following 2 lines of code X_train.append(training_set_scaled[i-60:i, 0]) &
y_train.append(training_set_scaled[i, 0]), we can see that we have a lag of 1 since we used i-60:i for X and i for y. So, for the training sets are periods of 60 days (e.g. see “ X_train[0,:].shape”) and y contains the next’days data of this 60 days period. Another way to verify the above is using “training_set_scaled[60] == y_train[0]”, “training_set_scaled[61] == y_train[1]” and so on... Hope this helps.