possible goodness of fit

[babaq]

regression evaluation seems to me an important part of the fitting process. I am wandering if LsqFit should included some of statistics for goodness of fit (such as r2, R2 or adjusted R2)?

[babaq]

ok, after some research, apparently, r2 or adjr2 are not suitable for nonlinear regression. but there are several ways to evaluate goodness of fit for nonlinear regression, such as standard error of the regression, pseudo r2/adjr2(already defined in the StatsBase but not implemented in the LsqFit), information criterion(aic, bic, etc.), and correlation between data and model predictions.

does any of these worth including in the LsqFit for general reference?

[pkofod]

I know this is ages ago, but I don't expect to add any R2-ish measures.