Non linear regression offers a number of advantages over linear regression:

• Equations do not need to be linearised (linearising equations may be difficult / not possible)

• Linearisiation can lead to systematic errors such as biasing the result of the fit

• Avoids challenges of negative values being discarded when logarithms are taken or very small positive values being large negative values once they are "logged". 

Non-linear regression can be achieved using various software including the use of spreadsheets, specialist data analysis software or through programming methods such as Python. Brief guides are provided below for Python and Excel.

Python

The Python pages include a graph plotting script which can be used for both linear and non-linear regression. You can set the desired best fit function to be used in the code. In the fitting process the constants will be determined.

Excel

Excel can carry out non-linear regression using the Analysis Toolpak. Instructions for installing this addon are provided below.

• Guide to using solver

Templates are provided for carrying out non-linear regression for kinetic analysis. Two templates are provided, and the relevant template will need to be selected to suit the analysis being undertaken.

• Loss of reactants

• Formation of product