Prediction of Average Rice Prices at the Indonesian Wholesale Level Using the Least Square Method

Abstract

Accurate forecasting of staple commodity prices is critical for economic planning, market stabilization, and food security, particularly in countries like Indonesia where rice is a dietary mainstay. This study investigates the application of the Least Square Method (LSM) as a mathematical model for predicting wholesale rice prices in Indonesia, aiming to evaluate its effectiveness and accuracy in a real-world setting. Using monthly price data collected from Statistics Indonesia (Badan Pusat Statistik) covering the period from January 2022 to December 2023, the LSM was employed to identify a linear trend equation that could be used for forecasting future prices. The model parameters were calculated based on time-indexed historical data, and the trend equation was used to predict prices for the upcoming 12 months, from January to December 2024. To assess the model’s forecasting accuracy, the Mean Absolute Percentage Error (MAPE) was used as the primary evaluation metric. The analysis revealed a MAPE value of 2%, which indicates a highly accurate prediction according to standard interpretative scales. These results confirm that the Least Square Method is a valid and practical approach for time-series forecasting of rice prices in Indonesia. The study highlights the potential of LSM as a simple yet effective tool for supporting policy decisions and market interventions. However, it also notes that the linear model may not account for external variables such as seasonal variation, policy shifts, or supply chain disruptions, suggesting that future research could explore multivariate or non-linear approaches for improved forecasting robustness.