CTShower Testing - Mar. 16 2005

Mirror of the best candidate showers page.

Changed the HitLogLikelihood_Primitive function to get the probability of drawing I_event from a Poisson distribution around I_LDF*P_TDF. (Original function was simply I_event*P_TDF).

01-29-03 Better LDF fit.
03-03-03 Better LDF fit.
03-24-03 About the same.
04-19-03 Now fails.
05-13-03 Better LDF fit.
05-17-03 Different but not better.
06-14-03 Still fails.
06-18-03 Better LDF fit, but still outside the array.
06-27-03 Still fails.
08-17-03 About the same.
11-20-03 Better LDF fit.
12-26-03 Still fails.

Generally, when the only problem was a low LDF, this seems to solve it. There seems to be a tendancy, however, to push the energy up and move the core out of the array sometimes. Also this fit is very sensitive to non-hit sites at the edge of the shower.

Changed the IntensityLogLikelihood_Primitive function so that for 0 hits, it does an approximate integration from 0 to 0.5 of Poisson(I_LDF). This doesn't make a substantial difference to any of the showers.

01-29-03 Similar.
03-03-03 Similar.
03-24-03 Similar.
04-19-03 Still fails.
05-13-03 Similar.
05-17-03 Similar.
06-14-03 Still fails.
06-18-03 Similar.
06-27-03 Still fails.
08-17-03 Similar.
11-20-03 Similar.
12-26-03 Still fails.