MODEL KLASIFIKASI PENERIMA PROGRAM KARTU INDONESIA PINTAR MENGGUNAKAN METODE XGBOOST

Poverty is one of the major factors contributing to the low quality of education in Indonesia. The Smart Indonesia Program (Program Indonesia Pintar) is a cash assistance program distributed through the Smart Indonesia Card (Kartu Indonesia Pintar/KIP) to support students from economically disadvantaged families, ranging from elementary to higher education levels. This study aims to classify students who are eligible to receive the Smart Indonesia Card using the XGBoost method, with a case study conducted at SMPN 02 Kebun Tebu. The dataset used in this study consists of independent and dependent variables. The independent variables include father’s age, mother’s age, father’s education level, mother’s education level, father’s income, mother’s income, number of family dependents, and students’ academic average scores. The dependent variable is the eligibility status of KIP recipients as the target class. Two classification models were developed using data split ratios of 70:30 and 80:20. The model with a 70:30 data split achieved an accuracy of 0.9048, a precision of 0.9034, a recall of 0.9072, and an F1-score of 0.9053. Meanwhile, the model with an 80:20 data split demonstrated better performance, with an accuracy of 0.9167, a precision of 0.9149, a recall of 0.9189, and an F1-score of 0.9169. The optimal model obtained from this study can be utilized by schools to support policy decision-making in determining eligible Smart Indonesia Card recipients, ensuring that educational assistance is distributed accurately, adequately, and equitably