Thai musical instrument recognition and classification using C5.0 algorithm

Abstract

This paper proposes the machine learning for the recognition and classification of Thai musicalinstruments. A set of features extracted from 9 Thai musical instruments are investigated for contentbasedclassification of musical instrument timbres. The Linear Prediction Coding is used to producesimple matrices from complex sound waveforms for instrument recognition. These matrices arepresented to a recognition model of C5.0 algorithm. A thorough evaluation of the resulting performancesand comparative analysis with SVM and Back Propagation Model is presented. A C5.0 model showsan accuracy rate of 99.77% outperforms the other two models due to its robustness, time-efficient andflexibility of modeling algorithm