Many analyses do not have a continuous model of the likelihood as a function of their parameter of interest. Often times, high mass searches and similar analyses who do not have such a model need to resort to a simple solution: They create one workspace for every (Higgs mass) scenario. In that series of workspaces, they then fit the likelihood tot he observed data and ultimately retrieve a single number from each of these fits -- the likelihood value. Given the POI (higgs mass) values associated to these input workspaces, those values can then be combined into something resembling a proper likelihood scan.
+MultiWorkspaceLikelihoodScan {
+Run1 {
+cosa {
+ggHcosa000 { <cosa=0.> }
+ggHcosa005 { <cosa=.05> }
+ggHcosa006 { <cosa=.06> }
+ggHcosa008 { <cosa=.08> }
+ggHcosa010 { <cosa=.10> }
+ggHcosa012 { <cosa=.12> }
+ggHcosa015 { <cosa=.15> }
+ggHcosa020 { <cosa=.20> }
+ggHcosa025 { <cosa=.25> }
+ggHcosa030 { <cosa=.30> }
+ggHcosa035 { <cosa=.35> }
+ggHcosa040 { <cosa=.40> }
+ggHcosa045 { <cosa=.45> }
+ggHcosa050 { <cosa=.50> }
+ggHcosa055 { <cosa=.55> }
+ggHcosa060 { <cosa=.60> }
+ggHcosa065 { <cosa=.65> }
+ggHcosa070 { <cosa=.70> }
+ggHcosa075 { <cosa=.75> }
+ggHcosa080 { <cosa=.80> }
+ggHcosa085 { <cosa=.85> }
+ggHcosa090 { <cosa=.90> }
+ggHcosa095 { <cosa=.95> }
+ggHcosa100 { <cosa=1.0> }
<snapshot="default",data="obsData"> @ ?;
<fit.initParam.mu=1.1> @ ?;
}
}
}