Available in: Uplift DRF
Use this option to specify the number of bins for the calculation of Area Under Uplift Curve (AUUC).
To calculate AUUC for big data, the predictions are binned into histograms. Because of this feature, the results should be different compared to exact computation. To get a value closer to the exact AUUC calculation, a higher number of bins has to be set.
To define AUUC, binned predictions are sorted from largest to the smallest value. For every group, the cumulative sum of observations statistics is calculated. The resulting cumulative uplift is defined based on these statistics.
The default value is -1, which means 1000 bins are used. The value should be higher than zero and less than number of data rows.