IMPLEMENTATION OF MACHINE LEARNING METHOD (NAIVE BAYES) IN PREDICTING SOCIAL ASSISTANCE RECIPIENT TARGETS IN PANGADEGAN VILLAGE

Abstract

Extreme poverty is a significant challenge in Pangadegan Village, Rancakalong District, Sumedang Regency. The village government has implemented a social assistance (bansos) program to address this issue. However, distribution often misses the target due to inaccurate data, suboptimal verification, and lack of transparency. This study uses the Naive Bayes method, a machine learning algorithm, to predict the correct recipients of social assistance. The method is chosen for its ability to perform probabilistic classification with minimal training data. The dataset used comes from the Extreme Poverty Eradication Acceleration Targeting Data (P3KE) with 342 testing records. Experimental results show that the Naive Bayes model has an accuracy of 51.18%, with a confusion matrix indicating decent predictions in some categories. Although the results are not optimal, the study demonstrates that Naive Bayes has potential in helping the village government distribute social assistance more accurately. Improved implementation can be achieved through more comprehensive data collection and stricter validation, leading to more precise and effective outcomes