I have iso-thermal degradation data and I want find out a reaction model that actually explains the degradation process. To optimize m and n from Sestak–Berggren kinetic model I need actual form of linearity coefficient (r or r-square).
Thank you
The proper usage for regression factors is important. The value of the regression factor R2 is a comparative metric for one linear model versus another. It should not be used by itself to validate anything. It certainly cannot be used to "actually explain" anything. Using R2 as a metric to "prove" the integrity of just one model, let alone to prove anything, is a badly-informed practice that is unfortunately still propagated in the literature.
With any non-linear model, "optimization" as you term it is more robust when it is done using non-linear regression fitting.
You can use, the solver add-in tool available to minimize your error between the experimental data and the equation that you would like to fit.
As Jeffrey pointed out, non-linear regression is much better. You can use the solver add-in tool to do both linear and non-linear regression.
You can use R2, or several other error functions, ifts you choice. Also Chi Squar is also a reasonable parameter to minimize the error between experiment and fitted data.
- Badges
- Science topic
- Similar topics
- Chemistry
- Pharmacology
- Medical Pharmacology
- Clinical Pharmacology