-
Hello
|
Beta Was this translation helpful? Give feedback.
Answered by
janezd
Jan 12, 2021
Replies: 1 comment 1 reply
-
The latter. If you run cross validation, you get predictions from cross validation. The former -- building the model from some data and then use it to make predictions on this very same data -- would be very wrong. (The user can of course explicitly require using the same data for training and testing.) |
Beta Was this translation helpful? Give feedback.
1 reply
Answer selected by
ajdapretnar
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The latter. If you run cross validation, you get predictions from cross validation.
The former -- building the model from some data and then use it to make predictions on this very same data -- would be very wrong. (The user can of course explicitly require using the same data for training and testing.)