Quantitative Structure-Activity Relationship (QSAR) models relate chemical structures to observed biological activities. A model is defined by a feature set (e.g., ECFP6) and a learning algorithm (e.g., Random Forest). Select a dataset, features, and algorithm, then click Build QSAR. Training may take time. Performance metrics will appear in Trained Models below, and saved models are available in Cheminformatics → QSAR Predictor.