For model validation of curve fitting nonlinear model, it is okay if we validate using root means squared, RMSE and absolute average deviation, AAD? Before this I also applied the adjusted-R2 but I am not confident with that value.
Usually the nonlinear curve-fitting done based upon least-squares method, giving a system of nonlinear algebriac equations in coefficients of the fitted function. I think, first the solution of this nonlinear system had to be confirmed .... in case some iterative method is being used ... nonlinear systems have many solutions but in a sensible range window the converge results must not change substantially ... :)
thank you for your answer. i did the iteration since i want to know the value of fitting parameter. but for validation i am not sure which validation is suitable.
SOLVING NON-LINEAR CURVE, USE LEAST SQUARE METHOD, WHICH GIVES APPROXIMATE SOLUTION.
- Badges
- Science topic
- Similar topics
- Correction