<![CDATA[Inflation is one of the key economic indicators used to monitor economic conditions and stability in Indonesia. The utilization of inflation data can be developed not only to describe economic conditions but also to support prediction and early detection of potential changes in inflation. This study aims to develop an inflation prediction model for Indonesia using the Random Forest Regressor algorithm based on macroeconomic indicators, namely the BI Rate, the Rupiah exchange rate against the United States Dollar (USD/IDR), and world oil prices. The study used monthly secondary data from April 2016 to June 2026 obtained from the Central Bureau of Statistics (BPS), Bank Indonesia, and the World Bank. The research stages consisted of data preprocessing, Exploratory Data Analysis (EDA), lag feature construction, model development, and evaluation using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and the coefficient of determination (R²). Two modeling scenarios were implemented: Model A without lag features and Model B with one-period lag features. The results showed that Model B achieved better performance, with an MAE of 0.5109, RMSE of 0.7084, and R² of 0.4764, compared with Model A, which obtained an MAE of 1.5712, RMSE of 1.8152, and R² of −2.4382. Feature importance analysis indicated that Lag_Inflation was the most influential variable, contributing 88.99%. These results indicate that incorporating historical inflation information improves the predictive performance of the Random Forest model and provides a potential basis for developing an Early Warning System (EWS) to support inflation monitoring and policymaking in Indonesia.]]>
Copyrights © 2026
