Penerapan Data Mining Dengan Algoritma SMOTE dan Naive Bayes untuk Klasifikasi Kelayakan Penerimaan Bantuan Sosial pada Kantor Camat Medan Timur

Abstract

The East Medan Sub-district Office faces challenges in determining the eligibility of social assistance recipients due to data inaccuracy, inconsistent criteria, and inefficient manual selection methods. This study aims to address these issues through the application of the SMOTE and Naive Bayes algorithms. SMOTE is used to address data imbalance, while Naive Bayes is applied to classify based on probability. The research process begins with collecting data on social assistance recipients, which is then processed using the SMOTE algorithm to increase the number of data in the minority category. Next, the Naive Bayes algorithm is applied to classify recipient eligibility based on available features. The calculation results show that the combination of these two algorithms significantly improved classification accuracy. Based on the calculation, the highest value is 0.008 "Eligible", so the tested data is categorized as Eligible to receive Social Assistance. If expressed as a percentage, this value is equivalent to 0.008 × 100 = 0.75 or 75%. Furthermore, the system developed in this study is designed to facilitate implementation in government agencies. With the resulting transparency and efficiency, this system is able to assist the East Medan Sub-district Office in distributing social assistance more precisely