KLASIFIKASI PENERIMA BANTUAN SOSIAL PROGRAM KELUARGA HARAPAN (PKH) KECAMATAN GONDANGLEGI MENGGUNAKAN METODE NAÏVE BAYES

Abstract

The Family Hope Program (PKH) is a crucial social assistance initiative for poverty reduction in Indonesia; however, the misallocation of aid remains a primary challenge. This study aims to improve the accuracy of classifying PKH recipients in Gondanglegi District using the Naïve Bayes method, focusing on socioeconomic data. The research employs the Gaussian Naïve Bayes algorithm, utilizing data from 1,000 PKH recipients characterized by 18 attributes grouped into four aspects: health, education, social welfare, and poverty. The process encompasses data preparation, attribute selection, data splitting (using 80:20, 70:30, and 60:40 ratios), and the encoding of categorical variables via One-Hot Encoding. Model evaluation was conducted using a confusion matrix to calculate accuracy, precision, recall, and F1-score. The results indicate that the Gaussian Naïve Bayes model achieved the highest accuracy—96.5%—with an 80:20 data split ratio. The model proved effective in classifying PKH recipients with a high degree of precision, thereby supporting more targeted decision-making by local government authorities. Consequently, implementing this method can enhance the efficiency and effectiveness of social assistance distribution within the PKH program.