Most econometric analyses of patent data rely on regression methods using a parametric form of the predictor for modeling the dependence of the response in focus on given covariates. These methods often lack the capability of identifying non-linear relationships between dependent and independent...