In this project, we employed different dimension reduction methods, PCA, SVD, MIM, and CMIM, and compared the performance of RandomForest and kNN classifiers on a Turkish Music Emotion Dataset (https://www.kaggle.com/datasets/blaler/turkish-music-emotion-dataset).
The results show that RandomForest, with forward feature selection based on Mutual Information performs best in this task.
The file project.pdf contains the final report of the project.
The folder Project2 contains the runnable code.