A Comparative Study of XBORE and XBOREOPT Hybrid Models for Ore Production Forecasting in the Mining Industry

Mining industries worldwide are increasingly adopting machine learning tools to handle the vast amounts of data generated daily from mine systems. Few hybrid models have been developed to handle data from mine systems, and this study investigates the comparative performance of machine learning algorithms XBORE and XBOREOPT for predicting ore production in mining operations. The study explores the efficacy of hyperparameter optimization techniques in enhancing predictive accuracy, robustness, and computational efficiency, which are crucial for optimizing resource allocation and operational planning in mining contexts. Through empirical evaluation using real-world data, this research aims to provide insights into the practical benefits of hybrid modeling approaches in industrial predictive analytics. In terms of Mean Squared Error (MSE), XBORE scores 347.8001, whereas XBOREOPT achieves a lower value of 340.5477. Since MSE penalizes larger errors more heavily, the lower MSE of XBOREOPT suggests it is more effective at minimizing significant prediction errors. This study combines the viability of hybrid modeling with practical applications in the mining industry, focusing on enhancing predictive accuracy and operational efficiency.