PiscesLogoSmallerStill Error dialog

Top  Previous  Next

The Levenberg-Marquardt numerical method to solve for the nonlinear regression is the best method known, however, it can fail and may often do so when the Richards curve is fitted to data. Failure is normally caused by the choice of inappropriate guesses for the values of the parameters to be estimated.

 

This box is shown when Levenberg-Marquardt has failed to converge to a solution.

 

error window

 

Click on OK and either try some other starting values until a solution is achieved or use a different model. One way you may be able to identify better starting values is to look at the output for models that were successful.