Optimasi Model Gaussian Mixture Model (GMM) untuk Klasifikasi Genre Musik Berbasis Mel-Frequency Cepstral Coefficients (MFCC)

Abstrak

Music genre classification is an increasingly relevant field as the number of digital music collections increases. The main challenge in this classification is to effectively capture the acoustic characteristics of different genres. This research proposes an optimization of the Gaussian Mixture Model (GMM) model to improve the accuracy of music genre classification using the Mel-Frequency Cepstral Coefficients (MFCC) feature. The dataset used covers various genres such as rock, classical, and jazz. The feature extraction process is carried out through MFCC and continued by training the GMM model with an optimized number of components. The test results show that the combination of MFCC and optimized GMM is able to improve the classification performance compared to conventional approaches. This study contributes to the development of an efficient machine learning-based music classification system.