Dear Colleagues,
I used Presence only Prediction (Maxent) in ArcGISPro 2.9.3 to generate an Output Trained Raster that makes sense given the inputs and is consistent with output made using the external version of Maxent. The Output Prediction Raster, however, differs markedly from the OTR in predicting values of 1 for almost the entire study area. Has anyone experienced this issue?
My model using 27 environmental variable with spatial thinning applied to 307 input point features.
The model did generate a set of warning messages as below:
WARNING 110427: Prediction values for 3800566 of 3800566 features had to be extrapolated due to explanatory variable values outside of training data ranges.
WARNING 110427: Prediction values for 19446 of 377370 features had to be extrapolated due to explanatory variable values outside of training data ranges.
WARNING 110423: Prediction values for 1 of 1 features had to be extrapolated due to explanatory variable values outside of training data ranges.
WARNING 110427: Prediction values for 3800566 of 3800566 features had to be extrapolated due to explanatory variable values outside of training data ranges.
Perhaps that is a clue? Would greatly appreciate any thoughts. Thanks!
David
