Study on prediction model of TCH degradation by three -dimensional electrocatalysis based on XGBoost and MLP

The prediction of Tetracycline hydrochloride (TCH) degradation using artificial intelligence (AI) has garnered increasing interest due to its potential to significantly reduce experimental costs. Despite this interest, the methods and techniques in this area remain underdeveloped and require further research. As AI technology advances, an expanding array of techniques becomes available for application in degradation prediction. In this study, polyaniline (PANI) was selected as a three-dimensional particle electrode for the degradation of TCH. We utilized five input variables: initial concentration, voltage, electrolyte concentration, particle concentration, and time, with the absorbance after degradation serving as the output. By integrating extreme gradient boosting (XGB) and a multilayer perceptron (MLP), we successfully developed the XGB_MLP model. This model achieved a predictive accuracy (R 2 ) of 98.52 % and a root mean square error (RMSE) of 0.052. Furthermore, a degradation rate of 83.48 % was observed for 10 ppm of TCH after 120 min. The findings of this study underscore the capability of the XGB_MLP model to accurately predict the degradation rate of TCH through electrochemical degradation.