Ridge classification: Interpreting prediction

Question

I'm particularly concerned about the following problem when using ridge classification for predicting binary outcome

When I'm encoding the binary outcome as 1 and 0; my model accuracy is 0.6456
When I'm encoding the binary outcome as -1 and 1; my model accuracy is 0.3525

Does anyone know whether the binary outcome in ridge classification is to be made -1 and 1 or 1 and 0 and why?

Any lead is appreciated

Are you still using a threshold of 0.5 when using -1 and +1 labels (for which the threshold should be 0)? — Dikran Marsupial, Apr 24 '22 at 07:22
Thanks @DikranMarsupial for pointing this out. Indeed this was the label problem — Arijit, May 17 '22 at 06:08

Tim · Answer 1 · 2022-04-24T07:09:51.353

1

Are use sure that when changing the labels you didn't reverse them? Because both results you got are the same, just the second one got the labels in reverse, $1 - 0.65 = 0.35$.

Both encodings are correct since there are variations of logistic loss for both cases, but which one to choose depends on which your software uses.

edited Apr 24 '22 at 07:09

answered Apr 24 '22 at 06:45

Tim

138,066

Ridge classification: Interpreting prediction

1 Answers1