The F test to Compare Fitting EquationsUnfortunately, Origin cannot predetermine which fitting function to use. The best way to compare fits with the same fitting equation is to compare the chi square value. However, this is not the best comparison for fits with different fitting equations. The standard way to compare these is the F test. F is defined as the following: F = {(SS1-SS2) / (DF1-DF2)} / {SS2/DF2} where SS1 and SS2 are the Sum of the Squares of each fit and DF1 and DF2 are the Degrees of Freedom for each fit. An F value close to 1.0 will indicate that first fit is correct.
Although at this time Origin does not calculate the F value for consecutive fits, you can get all needed values to calculate F. After performing a fit the Sum of Squares is stored in NLSF.SSR and the Degrees of Freedom is stored in NLSF.DOF. You can access these values through the script window by typing the following,
NLSF.SSR= (return)
You can also create your own LabTalk scripts to store these values and then calculate the F values.