I am using SVM classifier to classify my audio data (instrument strokes). I am using LDA as a feature reduction technique to reduce feature dimensions. After applying classifier the results I am getting is very low (only 45% accuracy). With the same data without any feature reduction and using PCA I am able to get more than 90% accuracy. What may be the problem?

  • $\begingroup$ Are you evaluating performance using cross-validation? $\endgroup$ – jonnor Aug 27 '18 at 11:10
  • $\begingroup$ What are your input features to the classifier? $\endgroup$ – jonnor Aug 27 '18 at 11:12

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.