Error Analysis of Linearization Methods in Regression of Data for the Van Laar and Margules Equations

Research output: Contribution to journalArticlepeer-review

15 Scopus citations

Abstract

Linearization is used extensively in regression and model testing of experimental data. In this work, Margules and Van Laar equation parameters have been calculated, for one sample data set, using several different linearization methods. The resultant parameter values were compared with values obtained by nonlinear regression. The increase in absolute error and change of its distribution due to the transformation of the data were also analyzed. It is concluded that linearization may lead to inaccurate or even completely incorrect parameter values which do not describe the original data adequately. The commonly used statistical tests may not detect the inaccuracy of the calculated parameter values. The ultimate test is the accurate recovery of the original activity coefficient data.

Original languageEnglish
Pages (from-to)2820-2825
Number of pages6
JournalIndustrial & Engineering Chemistry Research
Volume32
Issue number11
DOIs
StatePublished - 1 Jan 1993

Fingerprint

Dive into the research topics of 'Error Analysis of Linearization Methods in Regression of Data for the Van Laar and Margules Equations'. Together they form a unique fingerprint.

Cite this