2 replies to this topic

[mjPK]

Hi 

 

I am looking for a goodness of fit measurement for my models. Does Phoenix have Standard Error of the Regression or Residual Standard Error of the Regression this calculation built in? If not what is the simplest way to add this metric as global measure that I would not have to re-code each time I add a new model. 

 

 

 

Thanks in advance,

 

M

[Simon Davis]

Matthew, I'm not sure I follow your question, the Overall sheet for Maximum-likelihood model has the objective function (-2LL) and also AIC and BIC, whereas theta has CV% for each parameter;
 

there are the older least squares regression models, both library and custom also available but i would not recommend these as less flexible and powerful than PML used in  Maximum-likelihood model

 

 Simon.

Attached Thumbnails

• 

[mjPK]

Thanks for the reply. 

 

I am looking to access or implement another measure of "goodness of fit" I understand R^2 is not appropriate for non-linear models so Standard Error of the Regression or Residual Standard Error of the Regression. I understand this is sort of redundant with measures like AIC or BIC, however, this value was requested as well.