Forecasting Poverty Ratios in Indonesia: A Time Series Modeling Approach

Hidayat, Muhammad Fadlan
Henryka, Diva Nabila
Citra, Lovina Anabelle
Permai, Syarifah Diana

Engineering, Mathematics and Computer Science Journal (EMACS)
Vol. 6 No. 3 (2024): EMACS

Poverty is one of the main problems still faced by Indonesia today. To help find the right solution, an annual prediction of the poverty rate in Indonesia is needed. This study uses data on the 'Ratio of the Number of Poor People in Indonesia per year from 1998 to 2023' obtained from data.worldbank.org. The prediction methods used in this study include the Naïve Model, Double Moving Average, Double Exponential Smoothing, ARIMA, Time Series Regression, and Neural Network, with a total of 26 models. Of the 26 models, only 19 models passed the model comparison stage. Based on the evaluation results using the RMSE, MAE, MAPE, and MDAE metrics, it was concluded that the NNETAR Neural Network model showed the best performance among the six methods used to predict the poverty ratio in Indonesia.