Is there a statistical method I could use to validate a model with point data.

Discussion created by armein01 on Nov 26, 2012
I have a model which provides max likelihood (0 to 100), and plot data which provides known occurences of the modelled data.
The points do not always, and more likely do not intersect the pixels in the model.
Can I statistically evaluate the likelihood of the points belonging to the modelled data?
I have thought about using ellipse to model the direction of the modelled data.

Thanks, for any input!