Application of an Optimized Machine Learning Model Based on Bayesian Algorithm in Predicting Heavy Metal Adsorption by Biochar

Biochar adsorption is one of the most effective methods for mitigating heavy metal pollution; however, it remains quite challenging for its limited adsorption capacity and selectivity. To address these issues and enhance the adsorption performance of biochar for heavy metals in soil, we proposed a novel approach based on machine learning models optimized by the Bayesian algorithm. This solution efficiently determines the best machine learning model parameters from an infinite design space, reducing development time dramatically. The optimized models have shown improved performance compared to previous machine learning models, with the random forest model exhibiting the best results (R2 = 0.998; RMSE = 0.027). Importance analysis of heavy metal adsorption efficiency has ranked the following factors in order of relevance: adsorption environment > biochar characteristics > preparation conditions. This study provides valuable insights into improving the adsorption capacity of biochar materials for heavy metal pollutants in the environment, where the development of machine learning models and the application of Bayesian optimization algorithms have been proven to be effective methods. These new machine learning models and optimization methods are expected to be widely used in the environmental field, helping to guide future biochar material synthesis and heavy metal pollution control efforts. These research findings provide new ideas and technical means for addressing significant environmental issues