CLUSTERING DATA MAHASISWA BERDASARKAN ASAL DAERAH DAN ASAL SEKOLAH MENGGUNAKAN SELF ORGANIZING MAP (SOM) (Studi Kasus Teknik Informatika Universitas Muhammadiyah Malang)

Timely graduation of students is one of the problems that is difficult for every university to overcome, as well as in the Department of Informatics, University of Muhammadiyah Malang. This problem must be addressed immediately considering the quality of students will affect an accreditation of higher education and majors. Therefore, it is necessary to analyze which regions have the potential for timely graduation of UMM Informatics Engineering students. This study uses the Self Organizing Map (SOM) algorithm for clustering in order to find areas and schools that have an estimated student graduation opportunity on time. The independent variables used are regional origin, school origin, GPA and study period. The best parameters to be implemented on graduation data in 2015-2018 are learning rate = 0.2, and epoch = 500 with the Mean Square Error (MSE) value from the cluster = 0.001548016321479858