Quantitive Structure-Kinetic Relationships (QSKR) are regression or classification models that relate a set of physio-chemical properties or molecular descriptors to the target binding kinetics of a set of compounds. QSKR regression models use regression techniques to relate a set of molecular descriptors (referred to as "predictor variables" in Machine Learning terminology) to the kinetic parameter ("response variable"), where the predicted parameter is a continuous value. QSKR classification models, on the other hand, relate the predictor variables to a categorical value (e.g. slow or fast, active or inactive) of the response variable.
Image credit: Gaurav Ganotra (HITS, Heidelberg, Germany).