KLASIFIKASI LAGU BERDASARKAN LIRIK BAHASA INDONESIA MENGGUNAKAN METODE SUPPORT VECTOR MACHINE

Abstract

A song has lyrics that sets it apart from other songs, because the lyrics are a set of words that make up a song. Song lyrics exhibit specific properties different from plain text documents – many lyrics are for example have different frequencies when compared to other text documents. Classification based on the Indonesian song lyrics still very rarely done. This study uses a Support Vector Machine (SVM) to classify Indonesian song based on lyric with genres Pop, Rock, Punk and Rap. In this study using SVM-light implementation that generates file model and file prediction from the results of the classification. Further, lyrics may use differ greatly in the length and complexity of the language used, which can be measured by some statistical features such as word or verse length, and the amount of repeating text. In this paper, we present results for musical lyric classification on a test collection in order to demonstrate our analysis. The highest accuracy obtained in this study using 50 training data amounted to 67.5%.

Key words classification, Indonesian song, musical lyric, Support Vector Machine