V + jets / t ¯ t + jets background estimation for stop quark searches Thierry Caebergs, Gr´ egory Hammad Mons University V + jets / t ¯ ghammad@cern.ch t background estimation 1 / 5
V+jets and tt+jets estimation Analysis code ported to common framework successfully Macro reading from AnaTree ready Started to test performances V + jets / t ¯ ghammad@cern.ch t background estimation 2 / 5
Performances Previous studies have shown that stable results can be achieved with only one fixed parameter : ǫ ′ mistag , mis-tagging for v-like events. Once full analysis on tracks, should study if uncert. on estimated b-tagging eff. smaller than the one provided by BTV-11-003. Otherwise, use ǫ btag as fixed parameter. For testing purposes ; ǫ ′ mistag used as fixed parameter value set to 0.179 (derived from MC simulations) V + jets / t ¯ ghammad@cern.ch t background estimation 3 / 5
Performances for 5000 pb − 1 For n = 4 jets, using t ¯ t + jets and W + jets only ; Constant Par. Value ------------- ------------ euds 1.7900e-01 Floating Par. InitialValue FinalValue (+HiError,-LoError) ------------- ------------ ---------------------------------- Ntt 2.4509e+04 2.4681e+04 (+3.65e+02,-3.61e+02) Nv 2.3431e+04 2.2822e+04 (+3.60e+02,-3.62e+02) eb 8.0000e-01 7.9570e-01 (+8.90e-03,-9.10e-03) eudsc 2.0000e-01 1.8046e-01 (+3.97e-03,-3.84e-03) Estimated b-tagging eff. compatible with value derived from data using t ¯ t + jets events (BTV-11-003) : ǫ btag = 0 . 79 Input values for N t ¯ t - like and N V - like are taken from MC (fit stability already studied). V + jets / t ¯ ghammad@cern.ch t background estimation 4 / 5
Future dev. Mis-tagging efficiency : Derived from DD methods : BTV-11-004 Needs to determine average jet pt : from MC simulations ? from a control sample with no b-tagged jet (V-like enriched)? Errors will be incorporated in MLE via gaussian constraint Bias on estimations : Use pseudo-exp. Code already in place. V + jets / t ¯ ghammad@cern.ch t background estimation 5 / 5
Recommend
More recommend
