ANALISIS PERBANDINGAN KINERJA ALGORITMA SVM DAN KNN PADA SISTEM PENGENALAN GENRE MUSIK

Nasrullah, Zharfan Onetian and , Maryam , S.Kom., M.Eng (2026) ANALISIS PERBANDINGAN KINERJA ALGORITMA SVM DAN KNN PADA SISTEM PENGENALAN GENRE MUSIK. Skripsi thesis, Universitas Muhammadiyah Surakarta.

	PDF (Naskah Publikasi) L200224280_Zharfan Onetian Nasrullah_Library UMS.pdf Restricted to Repository staff only Download (651kB)
	PDF (Surat Pernyataan Tidak Publikasi) L200224280_Surat Pernyataan Publikasi Perpustakaan.pdf Restricted to Repository staff only Download (651kB)

Abstract

This study aims to examine and compare the effectiveness of Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) algorithms within a music genre identification system by utilizing the GTZAN Music Genre Dataset. The feature extraction process, conducted using the Librosa library, yielded 76 numerical features for each sample, including Mel-Frequency Cepstral Coefficients (MFCC) as the most dominant feature. The system pipeline involves data normalization using StandardScaler, dimensionality reduction via Principal Component Analysis (PCA), and hyperparameter tuning using GridSearchCV with 10-fold cross-validation. Experimental findings on the test data demonstrate that the SVM algorithm with an RBF kernel successfully achieved an accuracy of 76.5%. Conversely, the KNN algorithm with optimal parameters of n_neighbors=7, weights=distance, and the Manhattan distance metric obtained an accuracy of 72.5%. This study concludes that SVM consistently outperforms KNN in classifying music genres, owing to its ability to effectively separate high-dimensional and non-linear audio features.

Item Type:	Thesis (Skripsi)
Uncontrolled Keywords:	music genre classification, Support Vector Machine, K-Nearest Neighbor, MFCC, machine learning
Subjects:	T Technology > Information Technology > Artificial Intelligence T Technology > Information Technology
Divisions:	Fakultas Komunikasi dan Informatika > S1 Teknik Informatika
Depositing User:	ZHARFAN ONETIAN NASRULLAH
Date Deposited:	06 Aug 2026 07:26
Last Modified:	06 Aug 2026 07:26
URI:	http://eprints.ums.ac.id/id/eprint/147014

Actions (login required)

View Item