0

I am trying to classify between two different categories a set of 2 dimensional points with a linear SVC.

To train the model I use a dataset with 64 samples from class A and 230 from class B. Is this allowed in a basic Support Vector Machine classification or should I try to balance the two classes?

Mastodon
  • 25
  • 5

1 Answers1

1

Is it "allowed"? Sure. Your model will focus more on class B than on class A, however. You can try assigning class weights to deal with this, so that the model will try harder to avoid misclassifying an A example than a B example, perhaps making the total weights of the classes equal; standard SVM software supports this easily.

The advice here is applicable to SVMs as well as linear regression (though in SVMs, more than the intercept can change).

Community
  • 1
Danica
  • 21,852
  • 1
  • 59
  • 115