Akademik Veri Yönetim Sistemi
Journal of the Faculty of Engineering and Architecture of Gazi University, cilt.32, sa.4, ss.1221-1232, 2017 (SCI-Expanded)
Computer science has become a popular reseach topic in musicology with the transfer of musical works to digital media. Musical works are used as data in scientific researches and the computational music field is developing rapidly with the work done in this area. Representing Western musical works in symbolic form is easier than Turkish musical works and as a result most of the studies in this area focus on Western Music. However, in the last few years there are some interesting studies on using data mining, machine learning and classification techniques on Turkish maqam system. This study represents an experimental work that uses machine learning to recognize the maqams of the 1261 Turkish musical works. These musical works are assumed to be obtained by note recognition from audio files. We developed a software for using the data in MusicXML format with machine learning. This software also adds four different derived variables to the original data set in order to incerase the performance of the machine learning process. As a result of the study, we observed the perfomance of the "Random Forest" algorithm as 89.7%.