I have a distribution map produced with only presence data. And there is a certain number of presence data that is in no way included in the model. How can I evaluate the compatibility of the presence data not included in the model I have with the predictive values corresponding to these points in the potential distribution map? So we can also think like this: I have two columns. The first column has only 1 values, the second column has the predictive values. Which method would be the best approach to examine the relationship between these two columns?