Hi,
I have a question about anomaly detection, please.
Why should we tune X_j?
Secondly, if we have to tune it, why not also do it in the training set, please?"
Thank you very much,
Sao Mai
Can you give the video title and time mark?
Yes, sure :)!
It is week 1.
Vidéo: Developping and evaluating an anomaly detection system.
Time: 5.42 (for the cross validation set)
The second time (for the test set). Time: 7.07
Thank you 
It seems to me that he meant to “tune” the list of features by throwing in more features or removing some features.
Thank you very much ! It 's clear
Thank you!
