Research Article
An Extreme Learning Approach for Electronic Music Classification
@INPROCEEDINGS{10.1007/978-3-030-14657-3_9, author={Jing Wang}, title={An Extreme Learning Approach for Electronic Music Classification}, proceedings={IoT as a Service. 4th EAI International Conference, IoTaaS 2018, Xi’an, China, November 17--18, 2018, Proceedings}, proceedings_a={IOTAAS}, year={2019}, month={3}, keywords={Electronic Music Classification Kernel Principal Component Analysis Extreme Learning Machine Feature Extraction}, doi={10.1007/978-3-030-14657-3_9} }- Jing Wang
Year: 2019
An Extreme Learning Approach for Electronic Music Classification
IOTAAS
Springer
DOI: 10.1007/978-3-030-14657-3_9
Abstract
In order to recognize different kinds of electronic music, an extreme learning based method is proposed. Firstly, the feature of different electronic music data are extracted from cepstrum coefficient. Secondly, the kernel principal component analysis is adopted to reduce the dimension of features. Thirdly, in order to select appropriate parameters for an extreme learning machine, the genetic algorithm is used. Finally, experiments are carried out to verify the performance of the proposed electronic music classification method. In the experiments, we have established a database including four kinds of electronic music, i.e., “Guzheng”, “Lute”, “Flute”, and “Harp”. The experimental results show that the classification accuracy of the proposed method can reach up to 96%, while the wrong classification rate of the proposed method is only 14% which is much lower than existing electronic music classification models.