There are no reviews yet. Be the first to send feedback to the community and the maintainers!
Repository Details
- The purpose of this research is to compare and contrast the different algorithm’s results over the IRMAS dataset. The dataset contains 4 different instruments Electric Guitar, Acoustic Guitar, Violin and Trumpet. The features extraction process is performed using the Librosa library and numerous features were extracted such as tempo, beat frames, mfcc etc. Feature selection was done by dimension reduction techniques such as Principal Component Analysis and L1 based feature selection. The method used to classify the instruments were a series of five classification algorithms(KNN, Naive Bayes, Random Forest, Decision Tree, Bagging) using three different approaches which consists of K-folds, Hold out and Leave one out. The results were then compared and contrasted with respect to accuracy.
