<![CDATA[Food commodities are essential in developing countries such as Indonesia, and the government regulates food commodity prices in every province. However, price instability issues persist in certain provinces, creating challenges for effective policy control. Data science and statistical techniques play an important role in supporting the government’s efforts to monitor and manage food commodity prices. This study proposes the Stackelberg-K-Means method to predict the commodity price index in East Java. The proposed method is a collaborative framework that combines cluster analysis and stacking ensemble learning for time-series prediction. Cluster analysis is conducted first using Dynamic Time Warping as the distance measure, which is suitable for time-series data, resulting in two clusters for each commodity: rice, oil, and flour. The stacking model consists of base learners and a meta-learner. The base learner models include Ridge Regression, Random Forest, and Support Vector Regression, while the meta-learner uses Light Gradient Boosting. Parameter optimization is performed using grid search, and the proposed method is evaluated against AutoARIMA implemented in Python using both training and testing data. The results show that the proposed method outperforms the ARIMA model across all three error metrics: MAPE, MAE, and RMSE. For flour commodities, the scores are 0.042% versus 0.328%, 4.715 versus 37.57, and 6.34 versus 523.99, respectively. For rice commodities, the scores are 0.261% compared to 0.392%, 31.585 compared to 48.142, and 41.92 compared to 56.068. For oil commodities, the scores are 0.185% compared to 0.250%, 33.02 compared to 47.571, and 39.35 compared to 56.060.]]>
Copyrights © 2026
