The ability to analyse the contributions of specific spectral bins in a fit is very important. In limit setting, this can tell you which part of the signal spectrum is giving you the most sensitivity to the given signal decay. In general fitting this can tell you which part of a spectrum is contributing most to the fit.
The Summary and FitResults classes in #114 currently don't handle these per-bin values very well.
The added complication here is that penalty term values should not apply to single bins, as they apply only to the test statistic value summed over all spectral bins. This means that treatment of the test-statistics and penalty term values will need to be somewhat de-coupled, but, as much as possible. we want to try and retain the current functionality the Summary and FitResults classes provide.
The ability to analyse the contributions of specific spectral bins in a fit is very important. In limit setting, this can tell you which part of the signal spectrum is giving you the most sensitivity to the given signal decay. In general fitting this can tell you which part of a spectrum is contributing most to the fit.
The
SummaryandFitResultsclasses in #114 currently don't handle these per-bin values very well.The added complication here is that penalty term values should not apply to single bins, as they apply only to the test statistic value summed over all spectral bins. This means that treatment of the test-statistics and penalty term values will need to be somewhat de-coupled, but, as much as possible. we want to try and retain the current functionality the
SummaryandFitResultsclasses provide.