ESTIMATION OF ARRHENIUS MODEL PARAMETERS USING THREE LEAST SQUARES METHODS

STEPHEN G. HARALAMPU*, ISRAEL SAGUY, MARCUS KAREL

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

89 Scopus citations

Abstract

The effectiveness of three least squares regression methods has been assessed for the estimation of Arrhenius model parameters. The performance of each method was assessed by comparing the relative sizes and positions of the 90% confidence regions for the parameter estimates and the degradation predictions. The traditional regression scheme, Method I, of sequentially regressing concentration data to obtain rates followed by a regression of the rates versus temperature to obtain the Arrhenius parameters was shown to be the least desirable method. Method II was similar to Method I, but used a multiple linear regression to regress the concentration data through a single co value. This provided more precise estimates, but a bias was indicated due to the fact that the individual rate estimates were not independent. Method III utilized a nonlinear regression which did not evaluate the individual rates. This method was less precise than Method II, but showed no bias. Method III was superior to Method I in every respect.

Original languageEnglish
Pages (from-to)129-143
Number of pages15
JournalJournal of Food Processing and Preservation
Volume9
Issue number3
DOIs
StatePublished - Sep 1985
Externally publishedYes

Fingerprint

Dive into the research topics of 'ESTIMATION OF ARRHENIUS MODEL PARAMETERS USING THREE LEAST SQUARES METHODS'. Together they form a unique fingerprint.

Cite this