Generalized Kinetic Regression Analysis: Hypergeometric Kinetics

Abstract

A number of classical kinetic systems can be represented as matrix exponential functions. A generalization to matrix hypergeometric functions is presented. There arise from a generalized invariance principle or from an averaging of the exponential parameters over multivariate distributions. Matrix methods of handling these functions are described and illustrated. An estimation procedure using the "Gaussian iterant" is presented. The techniques are applied to the rehogram problem in physiology. The techniques may be used for either theoretical model- building or for empirical curve-fitting.