PENERAPAN MODEL ALGORITMA XGBOOST DALAM PENENTUAN STATUS PENERIMA SUBSIDI LISTRIK DI PT PLN (PERSERO)

Abstract

The distribution of electricity subsidies often experiences mis-targeting, where recipients come from capable groups. This study aims to implement the XGBoost (Extreme Gradient Boosting) algorithm to determine the eligibility status of electricity subsidy recipients study case at PT. PLN (Persero) ULP Masamba. The research method used is a computational experimental approach with machine learning-based modeling. The data includes variables such as Electricity Capacity, Rates, Description of Needs, and Matching Dukcapil.. The evaluation results of the model show an accuracy of 84.30%, recall of 85.52%, and AUC-ROC of 91.82%. These findings indicate that XGBoost is effective in classifying electricity subsidy eligibility objectively and fairly.