You apply PCA to reduce 100 features to 10 components before training an SVM. Your test accuracy improves. What is the most likely reason?

Question

Answer

PCA always improves classification accuracy regardless of dataset characteristics.

Answer

PCA labels the data automatically, improving the SVM's discriminative ability.

Answer

PCA adds new synthetic features that contain information not present in the original data.

Answer

PCA removes noisy or redundant features, reduces overfitting, and speeds up training; the retained components capture most signal with less noise.

Crowdly