Analisis Peramalan Indeks Harga Konsumen (IHK) Umum Provinsi DKI Jakarta dengan Pendekatan Autoregressive Integrated Moving Average (ARIMA)

Abstrak

This study aims to model and forecast the Consumer Price Index (CPI) of DKI Jakarta Province using the Autoregressive Integrated Moving Average (ARIMA) method. The data used were monthly CPI data from January 2015 to December 2019 obtained from Statistics Indonesia. The analysis was conducted through several stages, including data visualization, stationarity testing, model identification using ACF and PACF plots, parameter estimation, diagnostic testing, forecasting, and model accuracy evaluation. The results showed that the data were initially non-stationary and required transformation and differencing before modeling. Several ARIMA candidate models were generated, and ARIMA (2,1,2) was selected as the best model based on the lowest AICc value of 79.08. Diagnostic tests indicated that the residuals satisfied the assumptions of normality and white noise. Forecasting results suggested that the CPI would remain relatively stable with minor fluctuations in future periods. However, the model produced MAPE values of 68.75% for training data and 100.65% for testing data, indicating low forecasting accuracy. Therefore, although ARIMA (2,1,2) met the statistical assumptions, its forecasting performance for DKI Jakarta CPI data was still not optimal.