Comparative Analysis of ARIMA and Prophet Methods in Forecasting Producer-Level Medium Rice Prices in East Java Province

Abstract

Price instability of medium rice has important implications for food security and agricultural policy in Indonesia, making accurate price forecasting essential for decision-making. This study investigates the forecasting capability of the Autoregressive Integrated Moving Average (ARIMA) and Prophet models for predicting producer-level medium rice prices in East Java Province. Monthly producer price data spanning January 2020 to December 2025 were collected from the East Java Open Data Portal. The research followed the Cross-Industry Standard Process for Data Mining (CRISP-DM), covering problem identification, data exploration, preprocessing, model development, performance assessment, and forecasting deployment. Forecast accuracy was measured using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). Experimental results reveal that ARIMA(1,1,1) achieved the highest predictive accuracy, recording an MAE of 238.37, an RMSE of 293.11, and a MAPE of 1.88%, whereas Prophet produced larger prediction errors, with values of 2,062.14, 2,113.86, and 16.46%, respectively. Based on these results, the ARIMA model was selected to estimate producer-level medium rice prices for the 2026 forecasting period. Overall, the study demonstrates that ARIMA is better suited to the characteristics of the analyzed dataset and has the potential to support data-driven agricultural price monitoring and food price stabilization policies.

Keyword: ARIMA, Prophet, Time Series Forecasting, Medium Rice, Producer Price, East Java.