To address the problem of low prediction accuracy and large noise data in gold leaching rate prediction, a high-efficiency intelligent prediction method was proposed. Combining the improved XGBoost and LightGBM models, a dynamic learning rate, regularization optimization, and Bayesian optimization algorithm were used to design an ensemble learning model for gold leaching rate prediction. The results show that the ensemble learning model has a significantly improved prediction accuracy compared to traditional single models, and its MSE is about 28.8% and 22.9% lower than XGBoost and LightGBM, respectively. The model has high stability in real production environments. The ensemble learning model can effectively improve the prediction accuracy of gold leaching rate and has strong practical application value.

gold leaching rate  /  prediction  /  ensemble learning  /  XGBoost model  /  LightGBM model  /  numerical simulation