PERAMALAN PENGGUNAAN LISTRIK DI PROVINSI BALI MENGGUNAKAN METODE ARIMA

Abstract

This study aims to forecast electricity consumption in the Province of Bali using the ARIMA (Autoregressive Integrated Moving Average) method. The forecasting process is based on monthly electricity usage data spanning from January 2015 to June 2024. The initial analysis revealed a significant upward trend, with a notable decline in usage during 2020, coinciding with the COVID-19 pandemic. To address the issue of non-stationarity in the data, a differencing process was applied until stationarity was achieved, as confirmed by the Augmented Dickey-Fuller (ADF) test. Model identification was conducted using ACF and PACF plots, and several ARIMA models were evaluated based on their Akaike Information Criterion (AIC) values. The ARIMA(0,1,1) model was selected as the most suitable model due to its lowest AIC value and its compliance with diagnostic assumptions, including uncorrelated residuals (verified by the Ljung-Box test) and normally distributed residuals (confirmed by the Shapiro-Wilk test). The forecasting results demonstrated that the selected model provides stable predictions for the subsequent 12 months. This study is expected to contribute to effective planning and management of electricity demand in the Bali region.