This page provides implementations of the algorithm for computing a p-value that measures the fit of class labels to given probabilities, as outlined in the paper Testing the Calibration of Classification Models from First Principles by Stephan Dreiseitl and Melanie Osl.
We provide the download as one zip file, please extract and load the desired version as indicated below.
R version (calfp.R): load into R with "source('calfp.R')" Matlab version (calfp.m): place on Matlab's search path Mathematica version (CalFP.m): load into Mathematica with "Needs["CalFP`"]"