Penerapan JST Backpropagation untuk Prediksi Siswa Penerima Bantuan

Abstract

The role of education is to educate students and build quality human resources. School Operational Assistance (BOS) from the government is used to relieve students, not free schools but free students from school fees and extracurriculars. BOS can be used to provide scholarships to underprivileged students. Students are entitled to a scholarship if they meet three conditions, namely not having a Smart Indonesia Card (KIP), not having a Prosperous Family Card (KKS), and parental income of less than one million rupiah per month. The determination of BOS scholarship recipients is based on an agreement and joint decision by the management team, the teacher council, and the school committee. The process of determining students as scholarship recipients often encounters obstacles because it is done manually. This study applies backpropagation ANN to determine students who are eligible for BOS scholarships. The system is built based on the web using bootstrap as a framework, Javascript, and PHP. The training process uses a sample of 78 students with input architecture 4, hidden 6, output 1, learning rate 0.7, momentum 0.9, SSE 0.1, output 1, and a maximum of 10,000 iterations, an error of 0.03 is obtained. Testing with 21 student data obtained an error of 0.414923 and the accuracy of the test was 93.67%. The test results using the cause-effect graph technique can display the training process, graphs, error results, weight values and testing time with 93.67% success.