Penerapan K-Medoids dalam Penggerombolan Provinsi di Indonesia berdasarkan Indikator Pendidikan Jenjang SMA

Abstract

Program wajib belajar 12 tahun dicanangkan pemerintah pusat pada tahun 2012, tetapi di Indonesia masih terdapat ketimpangan pendidikan khususnya jenjang SMA. Untuk mengatasi hal tersebut, penting mengetahui daerah-daerah mana yang harus ditingkatkan pendidikannya. Penelitian ini bertujuan untuk menggerombolkan provinsi di Indonesia berdasarkan indikator pendidikan jenjang SMA serta mendapatkan deskripsi dari masing-masing gerombol yang terbentuk Metode yang digunakan pada penelitian ini adalah analisis gerombol. Oleh karena terdapatnya pencilan pada data, maka metode analisis gerombol yang tepat digunakan adalah K-Medoids. Hasil yang diperoleh dari penelitian ini adalah terbentuknya empat gerombol. Gerombol 1 terdiri dari tiga provinsi yang memiliki karakteristik tingginya persentase putus sekolah, rasio murid-guru, dan persentase ruang kelas kondisi baik, serta rendahnya persentase mengulang. Gerombol 2 beranggotakan 21 provinsi memiliki karakteristik tingginya persentase lulusan, rasio murid-sekolah, dan rasio murid-kelas sangat tinggi, serta rendahnya persentase putus sekolah. Gerombol 3 beranggotakan delapan provinsi memiliki karakteristik tingginya persentase guru layak, dan rendahnya rasio murid-guru, serta rasio murid-kelas. Sementara itu, Gerombol 4 beranggotakan dua provinsi memiliki karaktetistik tingginya persentase mengulang dan rendahnya persentase lulusan, rasio murid-sekolah, persentase guru layak, serta persentase kelas kondisi baik. Daerah yang perlu lebih diperhatikan ada pada Gerombol 4 karena memiliki beberapa aspek pendidikan yang kurang baik.

In 2012, the 12-year compulsory education program was launched by Indonesian government. But there is still educational inequality, especially at senior high school. To resolve this problem, it is important to know which areas should be improved in terms of education. This study clusters provinces in Indonesia based on senior high school education indicators and obtain descriptions of each cluster. The method used in this study is cluster analysis. Due to the presence of outliers, the clustering method K-Medoids is used. The result of this study is four clusters are formed. Cluster 1 has three provinces which are characterized by high percentage of dropout, student-teacher ratio, and percentage of classrooms with good condition, as well as low percentage of repetition. Cluster 2 has 21 provinces which are characterized by high percentage of graduate, student-school ratio, and student-classroom ratio, as well as low percentage of dropout. Cluster 3 has eight provinces which are characterized by high percentage of eligible teachers, as well as low student-teacher ratio and student-classroom ratio. Meanwhile Cluster 4 has two provinces which are characterized by high percentage of repetition, as well as low percentage of graduate, student-school ratio, percentage of eligible teachers, and percentage of classrooms with good condition. Areas in Cluster 4 should be paid more attention because they have some educational aspects that are not good enough.