Interesting model you are using (if two components are additive, it is a bi-phasic curve otherwise it will be a single peak curve). If x-values are replicated then this curve is describing the 'mean' of those replicates at each position of x-axis. So the mean-values-deviations from the fitted curve are the lack of fit. plot y-values versus the fitted values or residuals versus the x-axis values and see if under- or over-fitting issues. The two rates (k1=1/t1 and k2=1/t2) might be informative too. Say what 'y' and 'x' values are if needing more comments. Good luck...
Dear Mewa Singh Dhanoa,
Iam currently using the above equation to fit my data in origin . The fit is showing converged and succeeded , but the R2 values of the same is not 0.9 so Iam unable to understand my goodness of fit
I may have to see fitted curve plot...
>>> in the mean time shall we rule out pseudo degrees of freedom problem (I uploaded the following a while ago):
>>>
Beware of Pseudo Degrees of Freedom!
Source of Pseudo Degrees of Freedom:
►y-variable Replicates (i. e. multiple y-values for each value of the x-variable). Make sure you are using paired (1-for-1) single/average values of your y- & x-variables. Any Factor not part of analysis, average it out (really all factors of the design should be part of analysis unless you are analysing sub-part of a larger design.
►Any Repeated Measurements (e.g. repeated measurements over time of your design set up). Use appropriate methodology for their analysis (1) model out repeated observations and then analyse appropriate parameter(s) and any meaningful function(s) of the fitted parameters or (2) repeat observation ANOVA or MANOVA/REML based analysis.
►Average out precision or laboratory replicates prior to any analysis.
In any experiment you do, define clearly:
- Badges
- Science topic