Data and Scripts for 'Optimizing Protein–Ligand Docking Through Machine Learning: Algorithm Selection with AutoDock Vina

The data and scripts supporting this study include multioutput_rf_docking_performance.py, which handles multi-output Random Forest regression for docking score evaluation, and feature_performance_crossval.py, which assesses feature sets via cross-validation and ranks them by Mean Squared Error (MSE). feature_performance_gini_importance.py computes Gini importance and filters features, while multi_model_regression_performance.py compares various regression models (Ridge, Lasso, RandomForest, GradientBoosting, and XGBoost) on docking scores. Additional scripts, such as top_54_fingerprint_features.py and top_9_descriptor_features.py, identify the most important features from descriptors and fingerprints, combining top features for comprehensive analysis. Results are saved in Excel files for reproducibility. For more information, see the ReadMe file..