rememer how this algorithm works - its something along these lines... Fits the points in the bin with a flat line, then fits it with a step fuction (that has 3 free parameters, height before step height after step, and step point ). The step point that maximises the likelihood is chosen as the split point.
rememer how this algorithm works - its something along these lines... follows the same proceedure as HyperBinningMakerLikelihood but only the dimension is picked using the likelihood method. The splitting is done using the SMART method (HyperBinningMakerSmart) .