A feature selection algorithm of music genre classification based on ReliefF and SFS

Abstract

The feature selection algorithm can select the optimal feature subset from lots of music features about music genre classification, removing the irrelevant and redundant features to reduce the number of features and improving the accuracy and performance of classification. A feature selection algorithm based on ReliefF and Sequential Forward Selection (SFS), which is called “ReliefF-SFS algorithm”, is proposed in this paper. It can improve the ReliefF and SFS - ReiefF to remove some features of low weights but owning better classification results combined with other features, and SFS to have low performance, being not suitable for large datasets processing. The experimental results show that the ReliefF-SFS algorithm can remove the irrelevant and redundant features effectively, simplify the classification model, speed up training and improve the classification accuracy. In addition, it is of high performance.

DOI: 10.1109/ICIS.2015.7166651

9 Figures and Tables

Cite this paper

@article{Wu2015AFS, title={A feature selection algorithm of music genre classification based on ReliefF and SFS}, author={Meimei Wu and Yongbin Wang}, journal={2015 IEEE/ACIS 14th International Conference on Computer and Information Science (ICIS)}, year={2015}, pages={539-544} }