Evaluation of parameters for nonlinear thermodynamic models

Abstract A method based on the maximum likelihood principle has been developed for the determination of model parameters from experimental data when all the measured variables are subject to error. In addition to the best estimates of the parameters, this method also yields information useful in selection of appropriate models and evaluation of the accuracy of the data. Application of the method is illustrated in the reduction of binary vapor‐liquid equilibrium data.