ESTIMATION OF RATE CONSTANTS FROM MULTIRESPONSE KINETIC DATA

Alternative procedures for the estimation of several parameters from multivariate observations are evaluated for a complex system. Primary attention is given to methods based on generalization of nonlinear least squares and Bayesian estimation as proposed by Box and Draper. The latter method provides an effective means for estimation even in very complicated cases, and demonstrates a parametric sensitivity which allows a precision of estimation not possible with other techniques. © 1968, American Chemical Society. All rights reserved.