Journal: RSC Advances
Article Title: Comparative immobilization of labile Pb and Zn fractions in contaminated soil using jackfruit seed-, sugarcane bagasse-, and taro stem-derived biochars: a machine learning-assisted mechanistic elucidation
doi: 10.1039/d6ra01444e
Figure Lengend Snippet: Interpretability analysis of the exchangeable fraction (F1_Pb) of Pb and Zn (F1_Zn) using SHapley Additive exPlanations (SHAP) and partial dependence plots (PDP) based on random forest (RF) and extreme gradient boosting (XGB) models. SHAP plots illustrate both the magnitude and direction of the effects of input variables (pH, organic carbon (OC), electrical conductivity (EC), and biochar application rate) on model predictions, whereas PDPs describe the nonlinear relationships between the most influential variables and the predicted F1 metal concentrations.
Article Snippet: The suspensions were gently stirred and allowed to equilibrate before measurement. pH was measured with a Hanna HI 9124 pH meter (Hanna Instruments, Romania), while EC was measured with the built-in conductivity probe.
Techniques: