PENGELOMPOKKAN PERFORMA AKADEMIK MAHASISWA BERDASARKAN INDEKS PRESTASI MENGGUNAKAN K-MEANS CLUSTERING

Abstract— Indeks Prestasi (IP) or Grade Point Average (GPA)is a means that can be used as student performance indicator.Understanding student performance is expected to help academicinstitution to decide which treatment to be given to students. Inthis paper, K-Means Clustering algorithm is used to groupuniversity students academic performance into 4 clusters basedon students’ first four semesters IP. Each clusters has randominitial centroid. Centroid is checked in every iteration, until thereis no more change. Based on 50 experiments in grouping 100students academic performance data, the average error rate is0.92%. This shows that K-Means clustering is able to groupstudents academic performance consistentlyIntisari—Performa akademik mahasiswa berdasarkanperolehan nilai indeks prestasi (IP) digunakan sebagaiindikator untuk menyelesaikan masa perkuliahan mahasiswa.Pada penelitian ini algoritma K-Means Clustering digunakanuntuk mengelompokkan performa akademik mahasiwamenjadi 4 cluster. Setiap cluster memiliki centroid awal yangditentukan secara acak. Centroid mengalami perubahan jikasetiap proses clustering ada perpindahan data pada tiap clusterdan perubahan centroid berhenti apabila tidak ada lagiperubahan data. Berdasarkan 50 kali percobaan untukmengelompokkan 100 data, akurasi pengelompokkan performaakademik berdasarkan nilai rata rata standard error adalahsebesar 0,92%. Hasil ini menunjukkan pengelompokkanperforma akademik yang konsistenKata Kunci— K-Means Clustering, Grade Point Average, IndeksPrestas