"Bin Sizes for Lift Chart"
Rapidminer rookie here!
I'm attempting to create a lift chart from our Naive Bayes output and having a problem getting the output in the form I want.
Using the Create Lift Charts operator, I set the Number of Bins to "10" and the Binning Type to "Simple" and everything seems to work the way I'd expect. I get 10 bins and the range of my model confidence is the same for each bin. This creates bins with vastly different numbers of examples in each bin.
I want, instead, to create bins with the same number of examples in each bin. I used the binning type "Frequency", which is giving some strange results:
When I set the number of bins to 6, I get somewhat reasonable results: 6 bins with a similar number of examples in each.
Although, when I change the number of bins to 7 (or anything larger than that), I get a lift chart with about 70 bins.
Anyone have a similar experience?
All I want is equal sized bins!