A mathematical programming framework foroptimal model selection/validation of process data

摘要

This work considers the use of information indices for optimal model selection andvalidation of process data. The approach followed assumes the existence of a set offundamental process models associated with possible, although distinct, operatingregions. A 2-phase mathematical programming algorithm for the assessment ofstructural changes and optimal fitting of local models in data series is proposed. Thisapproach is used to determine the kinetic parameters of the gelation reaction of chitosanwith genipin, employing dynamical elastic modulus data.