Why is there a difference in the values of alphas and betas in the model after I calibrate?
Calibration finds the optimal values for a model's parameters.
As a result, the found values are dependent on the initial values, the tolerance setting, and the max number of iterations.
The key questions to ask of your found values are:
- Do the new values give the desired model?
- Is the model's fit (LLF/AIC, etc.) better or worse?