Could someone please explain how ROC works with SVM? Specifically i'm using RocCurveDisplay.from_predictions(y_test, y_pred, ax=ax[1]) which works fine.

Since the data for y_pred is not probabilities, but rather a vector that contains 0 or 1, is threshold constantly 0.5?



Your Answer

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