altenbach
Thanks for your advice! You advise that It should be very important to fit the data on the base of physical meanings , it is very useful!
But what I want to get is a strict description of goodness of fit in mathmatics.
Now I think mse is enough, In your nonlinear fit vi , why use [ (mse * covariance)^0.5 ]*fitted coefficient *100 to decibe the error?
Would you please tell me your math theory?
