随机森林回归 MAE

Question

I'm trying to create a random forest regression model to predict a continuous variable with the following characteristics (after log-transformation):

count 2868.000000 mean 4.566630 std 1.233496 min -0.356675 25% 3.745966 50% 4.604670 75% 5.452325 max 8.141889

The best result I get in terms of negative mean absolute error is -0.64. How should I evaluate this result? Is it good/decent/bad?

Thank you very much.

Answer 1

here is a link to the docs docs

the simple answer to your question is yes it is a good to have a -0.64 NMAE.

NMAE is the opposite of MAE but you can basically remove the minus and treat it as a normal MAE. so that means if you have a high MAE that means the NMAE is also high and that is not a good result but a low MAE can be seen as low NMAE and that's considered a good result of course depending on what you want to have.