Klasifikasi Gamelan Bali Menggunakan Support Vector Machine (SVM)

Abstract

Balinese gamelan is traditional Balinese music in the form of an orchestra which consists of various kinds of instruments, using pelog and celendro scales. Various aspects of life and communal activities in Bali cannot be separated from gamelan religious ceremonies, official ceremonies, to entertainment. For ordinary people and tourists, distinguishing Balinese gamelan is not an easy thing, because not everyone has the knowledge of Balinese gamelan, as well as the experience of hearing and seeing Balinese gamelan. This research aims to classify Balinese gamelan based on their functions into two classes, namely Ceremonial Gamelan and Gamelan for entertainment using Support Vector Machine (SVM), as well as Mel-Frequency Cepstral Coefficients (MFCC) for feature extraction from audio. The pre-processing stages include normalization and pre-emphasis, the features used are 13 MFCC coefficients, the classification process uses SVM with the RBF kernel. The results obtained is an SVM classification model with 86% accuracy using the parameters C=5, gamma=0.1, and tol=1e-1.