PREDICTION OF PIP RECIPIENTS USING K-NEAREST NEIGHBOR AT MI NURUL QOLBI

Education is a key foundation in the development of quality human resources. However, the rising cost of education makes some children unable to attend school due to their parents' financial limitations. To address this problem, the government launched the Indonesia Smart Program (PIP) which provides education funding assistance to eligible students. This research aims to develop an Information System that can predict the eligibility of students to receive PIP assistance using the K-Nearest Neighbors (KNN) algorithm. The data used comes from all students of Madrasah Ibtidaiyah (MI) Nurul Qolbi in the 2022-2023 school year. This research methodology involves testing with a value of k=13 and model evaluation is done using split ratio and cross-validation techniques. The results showed an accuracy of 98.98% from various split ratios (10:90, 20:80, 30:70, 40:60) and an accuracy of 99.24% using the 10-fold cross-validation technique. The accuracy results show excellent performance and provide important significance in the development of prediction systems to help the selection process of aid recipients more efficiently and reduce the administrative burden for schools. However, its application on a wider scale still requires further research, especially to test its consistency and effectiveness in different contexts and with more diverse datasets.