I have a 2d histogram and want to model it with a 2d keyspdf. The problem is that ~20% of the bins needs to be excluded because they occupy regions that are known behave differently (I have a 2d histogram that identifies each of the bins that need to be excluded).
So the goal is to have those bins be ignored in the fit and get the shape of the rest of the bins. I tried setting the weights of each of the to-be-ignored bins to a very small value when doing the “add” statements when I made the roodataset, but the fits were unchanged.
Is there either some way to get the points to have low weight, or even better, some less kludgey way to give roofit a list of bins to ignore?
Thanks very much for any thoughts on this.