.rds data files as simulated by the code in supplementary material 2 of "Calibrating machine learning approaches for probability estimation without calibration data" by Eleonora Di Carluccio, Georgios Koliopanos, Francisco M. Ojeda, Christian Weimar and Andreas Ziegler.. To access the data correctly, please download the folder, unzip it, and move the innermost rds-files folder to the Supplementary_material folder, replacing the empty rds-files folder. A description of the folder structure is shown below: |-----rds-files : folder where simulated data will get saved when the code is run| | before-recalibration : simulated data with probabilities before recalibration| | after-recalibration : simulated data with probabilities after recalibration| | after-recalibration-misspec : simulated data with probabilities after misspecified recalibration| File 06 in the Calibration_Sup2 folder will recalibrate the probabilities estimated from file 05 and save the results in rds-files:1) When setting both the parameters MISSPEC_MEAN and MISSPEC_PROP to FALSE, the correctly recalibrated data will be saved in folder 'rds-files/after-recalibration' (Simulation study 1).2) Setting the parameter MISSPEC_MEAN to TRUE and MISSPEC_PROP to FALSE will generate calibrations with misspecified mean values and save the results in the subfolder 'after-recalibration-misspec', all ending with the suffix 'mean-misspec' (Simulation study 2).3) Setting the parameter MISSPEC_PROP to TRUE and MISSPEC_MEAN to FALSE will generate calibrations with misspecified proportion values and save the results in the subfolder 'after-recalibration-misspec', all ending with the suffix 'prop-misspec' (Simulation study 3).4) Setting both the parameters MISSPEC_PROP and MISSPEC_MEAN to TRUE will generate calibrations with misspecified proportion and misspecified mean values and save the results in the subfolder 'after-recalibration-misspec', all ending with the suffix 'mean-prop-misspec' (Simulation study 4). Please note: computational time for the recalibration of all simulated datasets is quite extensive.Estimated run times are presented as comments in the parameter selection section of file 06 (lines 70-100).For this reason, we have uploaded a copy of the rds-files folder containing all the necessary data to Zenodo.