Henny Pramoedyo, Atiek Iriany, Marjono, Yani Quarta Mondiana
Accurate prediction of sugarcane production is vital for effective agricultural planning in Indonesia. However, traditional methods can be hampered by spatial variations and outliers in yield data. This study addresses this challenge by employing Geographically Weighted Panel Regression (GWPR) with a robust M-estimator to predict sugarcane production in East Java, Indonesia, from 2019 to 2022. Addressing the challenges posed by outliers in sugarcane yield data from regions like East Java. This research combines panel data analysis and geospatial regression to capture the intricate spatial and temporal dynamics of sugarcane production. Based on MAE the results demonstrate the significant improvement in predictive performance achieved through robust estimation within the GWPR model, emphasizing the effectiveness of this approach in refining sugarcane production forecasts in the Indonesian context. This study highlights the potential of robust estimation methods to enhance agricultural forecasting models in Indonesia's sugarcane industry. © 2025 NSP Natural Sciences Publishing Cor.
Department of Statistics, Brawijaya University, East Java, Indonesia; Department of Mathematics, Brawijaya University, East Java, Indonesia
