Timeline for Applying de-standardised ridge regression coefficients to new test data - how to best handle the mean of y_test?
Current License: CC BY-SA 4.0
2 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 10, 2020 at 5:38 | comment | added | user284388 | Thank you Matthew for your input. It sounds like you are implying one should act as though one doesn't know one's test data's y-values, as one wouldn't with actual new data. You then sugget that one should use MSE instead of R^2: However, I shall like to remark that I don't see how using MSE instead of R^2 frees me from intercept differences (MSE is not relying on the mean, but still intercept-sensitive). I was then thinking about using the correlation between predicted and observed values, what do you think about that? | |
| May 8, 2020 at 16:42 | history | answered | Matthew Drury | CC BY-SA 4.0 |