Generalized ridge analysis with application to population pharmacokinetics/dynamics

Given a set of measurements on a patient in clinical studies or on a drug response in bioassays and immunoassays, the data series for the individual can be incomplete, noisy, and haphazard. Hence, meaningful analysis on such limited data is at best difficult given the typical assumptions on the nonlinear structure of systematic and random components involving both individual and population effects. This paper describes a simple direct semiparametric procedure to incorporate population-wide information from as many individuals as required to support analysis of data on any specific individual. This notion is motivated by the pattern-processing capabilities of experts as they evaluate data on an individual based on their reservoir of accumulated knowledge of the given application on many individuals.