I implemented a Gaussian Naive Bayes classifier and I got a test score (99,99%) higher than the train score (96,87%) Is this normal or does it mean that my model is underfitting ? Thank you.

  • $\begingroup$ It could be any number of things, try a different split. $\endgroup$ Commented Dec 21, 2022 at 10:12
  • $\begingroup$ You should add more information so that we can help you better (rows quantity, type of data, etc.) $\endgroup$ Commented Dec 21, 2022 at 21:13

1 Answer 1


Higher performance on test data set relative to train data set can happen for many reasons. Mostly likely it is due to chance.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.