Effect of regression approach in the estimation of nonlinear model parameters on process design and simulation: applications to kinetic and thermodynamic models

An inside-variance estimation method (IVEM) for regression of the kinetic p arameters in kinetic models and binary interaction parameters in thermodyna mic models is proposed. This maximum likelihood method involves the re-comp utation of the variance for each iteration of the optimization procedure, a utomatically re-weighting the objective function. Once the objective functi on is selected, most regression strategies consist of weighting the objecti ve function by pre-selected values, usually based on experimental error est imates (i.e. standard deviation), converting the problem into a traditional weighted least squares minimization. A problem with the traditional approa ch is that the experimental error estimation from the maximum-likelihood re gression cannot be unbiased, without using replicates. Thus, the use of exp erimental variances to weight the objective function does not necessarily p roduce optimum parameters for prediction purposes, even if the values obtai ned represent the global minima of the objective function. The new method s ubstantially improves the model predictions when compared with traditional least square regression methods. (C) 2000 Elsevier Science Ltd. All rights reserved.