Detection of Indonesian Dangdut Music Genre with Foreign Music Genres Through Features Classification Using Deep Learning

A piece of music consists of elements that form a single unit that makes it sound beautiful. Music is not only used as entertainment but also as a variety of means. These kinds of music have a long history and journey so that humans can recognize the flow of music through rhythm, vocals, and supporting instrumentals. The art of music has changed along with the development of technology to influence the musical genre. This paper aims to detect the dangdut music genre among Western music genres. The dangdut music genre dataset used in this study uses its own dataset of 100 data, while the western music genre dataset uses the popular GTZAN dataset. The GTZAN dataset has 1000 data from 10 music genres. In comparison, the duration of the music used for each genre in this study uses 30 seconds. The method used in detecting the classification of these genres is deep learning. The accuracy obtained when using this dataset in deep learning modeling is above 90%. However, when the data of the dangdut music genre was added to the dataset, the accuracy decreased to around 76%. Several model tests have been carried out and always get the accuracy of the dangdut music genre of 88.89%. It proves that dangdut music is somewhat different from other foreign music genres, but some are still detected similar to jazz and pop music genres.