Hi I have a question regarding the pricing speed problem. As the second question need to select informative features, I wonder if we can select informative features from the whole data set then we train our model in each batch set. Or we have to do feature engineering based only on the first batch size set?
We would not comment on the way you obtain your model, that’s up to you to build the best model you can with the data that you have ! Good luck with that. Lionel.