Let's say that I have training data and test data. I trained the model with the training data and then constructed a confusion matrix of the predictions of both. I am getting that the prediction of training data is lower than the test data.

This my training data prediction:

enter image description here

This is the test prediction confusion matrix (bottom screen). What does it mean? Is my model bad because my training accuracy is lower than my test? What about my 95% CI which is lower in the test data than in the train data prediction? I do not really understand everything even after searching on Google. I am using linear SVM method.

If anyone knows something I would be happy to listen to it.

enter image description here

  • $\begingroup$ Both acc values are within the 95% CI range, so you cannot say that one is lower than the other. Your training acc (0.87) is in the CI-range of the test acc [0.81-0.96]. For interpretation of the results also see this post: stats.stackexchange.com/questions/253407/… $\endgroup$
    – Peter
    Feb 4 at 10:04

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.