I have a number of regression models from which I'm trying to choose the best performing one.

I have computed SSE, AIC and BIC for all, including distributions of errors from predictions on unseen data.

M1 80.629 344.78 328.26
M2 374.72 698.62 688.71
M3 80.625 339.46 326.25
M4 80.764 339.95 326.73
M5 80.609 350.01 330.19

Error qq: enter image description here

Any ideas on how to decide the best from this bunch based on these criteria? I understand these metrics should tell you the best performing model from the lowest values for each, but as seen in the table, the differences between a number of models are minimal. Apologies if the question is unclear or the solution obvious, Some of this is fairly foreign to me.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.