Penggunaan Metode Naive Bayes pada Klasifikasi Risiko Angka Putus Sekolah di Sumatera Utara

Abstract

The school dropout rate is a crucial indicator for assessing the quality and equity of education in a region. A high dropout rate often reflects structural issues such as poverty, limited access to educational facilities, and a shortage of teachers, which ultimately lead to a decline in the quality of human resources. This condition requires an analytical method capable of identifying patterns and accurately predicting the level of dropout risk. This study aims to classify the dropout risk level in North Sumatra Province using the Naive Bayes classification algorithm, an effective probabilistic model for categorical prediction based on numerical data. The model was built using several independent variables, namely the poverty rate, population density, number of schools, number of teachers, student-to-school ratio, student-to-teacher ratio, number of male and female students, and the school dropout rate. These variables were selected because they represent socio-educational factors that influence the continuity of education. The classification process was carried out using the Gaussian Naive Bayes approach, which assumes that each feature follows a normal distribution within each class, thus enabling probabilistic predictions with high efficiency. The results show that the model achieved a prediction accuracy of 87.88% and effectively classified regions into “High” and “Low” school dropout risk categories.