TCanvas* pHitSpecPos( )
{
TCanvas* c = new TCanvas("cHitSpecPos","cHitSpecPos",1000,1100);
c->Divide(2,2);
TVirtualPad* p; TH2D *h;
TText t; t.SetTextColor(4);
p =c->cd(1); p->SetLogy(); p->SetGrid(1,0);
h = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhi_BarMinus");
h->GetYaxis()->SetRange(4,33);
h->DrawCopy("box");
t.DrawTextNDC(0.17,0.15, "BARREL MU MINUS");
p =c->cd(2); p->SetLogy(); p->SetGrid(1,0);
h = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhi_BarPlus");
h->GetYaxis()->SetRange(4,32);
h->DrawCopy("box");
t.DrawTextNDC(0.17,0.15, "BARREL MU PLUS");
p =c->cd(3); p->SetLogy(); p->SetGrid(1,0);
h = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhi_EndMinus");
h->GetYaxis()->SetRange(4,32);
h->DrawCopy("box");
t.DrawTextNDC(0.17,0.15, "ENDCAP MU MINUS");
p =c->cd(4); p->SetLogy(); p->SetGrid(1,0);
h = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhi_EndPlus");
h->GetYaxis()->SetRange(4,32);
h->DrawCopy("box");
t.DrawTextNDC(0.17,0.15, "ENDCAP MU PLUS");
return c;
}
TCanvas* pHitSpecPosGen( )
{
TCanvas* c = new TCanvas("cHitSpecPosGen","cHitSpecPosGen",1200,600);
c->Divide(2,1);
TVirtualPad* p; TH2D *h; TH2D *hAtPhi0;
TText t; t.SetTextColor(4);
p =c->cd(1); p->SetLogy(); p->SetGrid(0,1);
h = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhiGen_Minus");
hAtPhi0 = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhiGenAtPhi0_Minus");
h->GetYaxis()->SetRange(4,33);
h->GetXaxis()->SetNdivisions(505); h->GetXaxis()->SetLabelSize(0.04);
h->DrawCopy("box");
hAtPhi0->SetLineColor(2);
hAtPhi0->DrawCopy("box same");
t.DrawTextNDC(0.17,0.15, "BARREL MU MINUS");
std::cout <<h->GetTitle() << std::endl;
for (int iy = 1; iy <=h->GetNbinsY(); iy++) {
std::cout <<" pt: " << h->GetYaxis()->GetBinLowEdge(iy);
double xmin=100.;
double xAtMax = 0.; double valAtMax = 0.;
double xmax=-100.;
for (int ix = 1; ix <=h->GetNbinsX(); ix++) {
double val = h->GetBinContent(ix,iy);
double xbmin = h->GetXaxis()->GetBinLowEdge(ix);
double xcent = h->GetXaxis()->GetBinCenter(ix);
double xbmax = h->GetXaxis()->GetBinUpEdge(ix);
if (val > valAtMax) { valAtMax = val; xAtMax = xcent; }
if (val > 1 && (xbmin < xmin)) xmin = xbmin;
if (val > 1 && (xbmax > xmax)) xmax = xbmax;
}
std::cout <<" set DPHI0 = "<<1.025-xAtMax
<<"; set DPHI_MARGIN = " <<std::max( (xAtMax-xmin), (xmax-xAtMax))
<<";" << std::endl;
}
p =c->cd(2); p->SetLogy(); p->SetGrid(1,1);
h = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhiGen_Plus");
hAtPhi0 = (TH2D*)gROOT->FindObject("hHitSpec_PtVsPhiGenAtPhi0_Plus");
h->GetYaxis()->SetRange(4,32);
h->GetXaxis()->SetNdivisions(505); h->GetXaxis()->SetLabelSize(0.04);
h->DrawCopy("box");
hAtPhi0->SetLineColor(2);
hAtPhi0->DrawCopy("box same");
t.DrawTextNDC(0.17,0.15, "BARREL MU PLUS");
std::cout <<h->GetTitle() << std::endl;
for (int iy = 1; iy <=h->GetNbinsY(); iy++) {
std::cout <<" pt: " << h->GetYaxis()->GetBinLowEdge(iy);
double xmin=100.;
double xAtMax = 0.; double valAtMax = 0.;
double xmax=-100.;
for (int ix = 1; ix <=h->GetNbinsX(); ix++) {
double val = h->GetBinContent(ix,iy);
double xbmin = h->GetXaxis()->GetBinLowEdge(ix);
double xcent = h->GetXaxis()->GetBinCenter(ix);
double xbmax = h->GetXaxis()->GetBinUpEdge(ix);
if (val > valAtMax) { valAtMax = val; xAtMax = xcent; }
if (val > 1 && (xbmin < xmin)) xmin = xbmin;
if (val > 1 && (xbmax > xmax)) xmax = xbmax;
}
std::cout <<" set DPHI0 = "<<1.025-xAtMax
<<"; set DPHI_MARGIN = " <<std::max( (xAtMax-xmin), (xmax-xAtMax))
<<";" << std::endl;
}
return c;
}
void EstimateBg_76X(bool save=0, std::string in = "",
std::string out = "/afs/cern.ch/user/j/jkarancs/public/NOTES/notes/AN-14-290/trunk/Plots/v1.0/", std::string ext="png") {
gStyle->SetOptTitle(0);
gStyle->SetOptStat(0);
bool latex = save;
bool ABCD_prime = 0;
bool TT_only = 0;
std::stringstream ss, ss2;
ss<<DPHI_CUT; ss2<<R_CUT;
std::string dphi_cut = ss.str().replace(ss.str().find("."),1,"p");
std::string r_cut = ss2.str().replace(ss2.str().find("."),1,"p");
std::string filename = in.size() ? in :
//"results/Plotter_out_2016_05_31_08h48m57_replot.root";
"results/Plotter_out_2016_06_24_14h28m51.root";
std::vector<std::string> samples[4];
//samples[0].push_back("TTJetsMGHT");
//samples[0].push_back("TTJetsMG");
//samples[0].push_back("TTJetsNLOFXFX");
//samples[0].push_back("TTNLO");
//samples[0].push_back("TTNLOHerwig");
//samples[0].push_back("TTPowheg");
//samples[0].push_back("TTPowhegmpiOff");
//samples[0].push_back("TTPowhegnoCR");
//samples[0].push_back("TTPowhegHerwig");
//+data+ samples[1].push_back("SingleElectron");
//+data+ samples[1].push_back("SingleMuon");
if (TT_only) {
samples[1].push_back("TTJetsMGHT");
samples[1].push_back("TTJetsMG");
samples[1].push_back("TTJetsNLOFXFX");
samples[1].push_back("TTNLO");
samples[1].push_back("TTNLOHerwig");
samples[1].push_back("TTPowheg");
samples[1].push_back("TTPowhegmpiOFF");
samples[1].push_back("TTPowhegnoCR");
samples[1].push_back("TTPowhegHerwig");
} else {
samples[1].push_back("TTJetsMGHT");
//samples[1].push_back("TTJetsMG");
//samples[1].push_back("TTJetsNLOFXFX");
//samples[1].push_back("TTNLO");
//samples[1].push_back("TTNLOHerwig");
//samples[1].push_back("TTPowheg");
//samples[1].push_back("TTPowhegmpiOff");
//samples[1].push_back("TTPowhegnoCR");
//samples[1].push_back("TTPowhegHerwig");
samples[1].push_back("ZJets");
samples[1].push_back("TTX");
samples[1].push_back("WJets");
samples[1].push_back("Diboson");
samples[1].push_back("Top");
samples[1].push_back("QCD");
//ZERO samples[1].push_back("TZQ");
//ZERO samples[1].push_back("ZJetsToQQ"); // Also WJetsToQQ
//ZERO samples[1].push_back("GJets");
}
//samples[1].push_back("Data");
// NTop Sideband All background summed
samples[2].push_back("All Bkg.");
// Signal in NTop bins
samples[3].push_back("T1tttt");
bool baderror = false;
double weight[] = { 0.32686, 0.0505037, 0.00921411, 6.80717, 0.354934, 0.00484915 };
int rebin = /*(R_CUT*10-int(R_CUT*10))==0 ? 10 :*/ (R_CUT*20-int(R_CUT*20))==0 ? 5 : (R_CUT*50-int(R_CUT*50))==0 ? 2 : 1;
double sideband_fit_low_range[] = { 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15, 0.15 };
int i_h_side[] = { 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
int i_h_signal[] = { 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 };
double scale_factors[] = { 1, 1, 1, 1, 1, 1, 1}; // All normal
//double scale_factors[] = { /* TT */ 1, /* W */ 1, /* Z */ 1, /* T */ 1, /* TTV */ 1, /* QCD */ 2, /* VV */ 1 }; // QCD high
//double scale_factors[] = { /* TT */ 5, /* W */ 1, /* Z */ 1, /* T */ 1, /* TTV */ 5, /* QCD */ 1, /* VV */ 1 }; // TT/TTV high
//double scale_factors[] = { /* TT */ 1, /* W */ 2, /* Z */ 2, /* T */ 2, /* TTV */ 1, /* QCD */ 1, /* VV */ 2 }; // T/V/VV high
Double_t Rranges_ABCD[][4] =
{ { DPHI_CUT, 3.2, 0.0, DPHI_CUT },
{ R_CUT_LOW-1e-10, R_CUT, R_CUT-1e-10, 1.20 },
{ R_CUT_LOW-1e-10, R_CUT, R_CUT-1e-10, 1.20 },
{ R_CUT_LOW-1e-10, R_CUT, R_CUT-1e-10, 1.20 } };
bool doFitting = false;
double sum_a = 0, sum_b = 0, sum_c = 0, sum_d = 0, sum_d_abcd = 0, sum_d_nevt = 0;
double sum_a_err = 0, sum_b_err = 0, sum_c_err = 0, sum_d_err = 0, sum_d_abcd_err = 0;
double sum_b_fit = 0, sum_d_fit = 0, sum_d_fit_comb = 0;
double sum_b_fit_err = 0, sum_d_fit_err = 0, sum_d_fit_comb_err = 0;
double comb_d = 0, comb_d_err = 0, comb_d_abcd = 0, comb_d_abcd_err = 0;
if (latex) {
printf("\\begin{table*}[htbH]\n");
printf("\\small\n");
printf("\\begin{center}\n");
printf("\\topcaption{Estimated Standard Model background yields in ABCD regions. A, B is in the sideband, C and D is the signal band, D is the signal region.\\label{tab:SMBkgEstimate}}\n");
printf("\\begin{tabular}{lrrrrrrrr}\n");
}
TFile *f = TFile::Open(filename.c_str());
for (size_t iMethod = 0; iMethod<4; ++iMethod) if (!(iMethod==0&&samples[0].size()==0)){
// Print Top row for each method
if (latex) {
if (iMethod==0) {
printf("\\hline\n");
printf("Method 2 & A & B & C & D = B*C/A & D obs. & Ratio pred./obs. & Pull & KS test\\\\\n");
} else if (iMethod==1){
printf("\\hline\n");
printf("Method 1 & A & B & C & D = B*C/A & D obs. & Ratio pred./obs. & Pull & KS test\\\\\n");
}
printf("\\hline\n");
} else {
std::stringstream r_sb_cut;
if (R_CUT_LOW==0) r_sb_cut<<"R<"<<R_CUT;
else r_sb_cut<<R_CUT_LOW<<"<R<"<<R_CUT;
const char* prime = ABCD_prime ? "'" : "";
//if (iMethod==0) printf("| *Sample* | *A (DPhi>2.8, SB)* | *B (DPhi<2.8, SB)* | *C (DPhi>2.8, Sig.B.)* | - | *D = B*C/A pred.* | *D (DPhi<2.8, Sig.B.) obs.* | *Ratio pred./obs.* |\n");
//else if (iMethod==1) printf("| *Sample* | *A (R<%1.1f, SB)* | *B (R>%1.1f, SB)* | *C (R<%1.1f, Sig.B.)* | *D = B (R fit, SB) * C/A pred.* | *D = B*C/A pred.* | *D (R>%1.1f, Sig.B.) obs.* | *Ratio pred./obs.* | \n", R_CUT, R_CUT, R_CUT, R_CUT);
if (iMethod==0) printf("| *Sample* | *A (DPhi>%1.1f, %s)* | *B (DPhi<%1.1f, %s)* | *C (DPhi>%1.1f, R>0.4)* | *D = B*C/A pred.* | *D (DPhi<%1.1f, R>0.4) obs.* | *Ratio pred./obs.* | *Pull (pred-obs)/error* | *KS test* |\n", DPHI_CUT, r_sb_cut.str().c_str(), DPHI_CUT, r_sb_cut.str().c_str(), DPHI_CUT, DPHI_CUT);
else if (iMethod==1) printf("| *Sample* | *A%s (%s, <2 tag)* | *B%s (R>%1.1f, <2 tag)* | *C%s (%s, 2 tag)* | *D%s = B%s*C%s/A%s pred.* | *D%s (R>%1.1f, 2 tag) obs.* | *Ratio pred./obs.* | *Pull (pred-obs)/error* | *KS test* |\n", prime, r_sb_cut.str().c_str(), prime, R_CUT, prime, r_sb_cut.str().c_str(), prime, prime, prime, prime, prime, R_CUT);
}
TH1D *h_side_sum, *h_signal_sum;
for (size_t iSample = 0; iSample<samples[iMethod].size(); ++iSample) {
std::string canname =
iMethod==0 ? std::string("DPhiBins")+(ABCD_prime ? "/RBins_0To1HadTop_" : "/RBins_2HadTop_")+samples[iMethod][iSample] :
iMethod==1 ? std::string("RFine/Tau32Cuts_")+(ABCD_prime ? "Fail" : "Pass")+"DPhiCut_"+samples[iMethod][iSample] :
iMethod==2 ? std::string("RFine/Tau32Cuts_")+(ABCD_prime ? "Fail" : "Pass")+"DPhiCut_Background" :
iMethod==3 ? std::string("RFine/Tau32Cuts_")+(ABCD_prime ? "Fail" : "Pass")+"DPhiCut_"+samples[iMethod][iSample] : "";
TCanvas *can = (TCanvas*)(f->Get(canname.c_str()));
can = (TCanvas*)can->Clone();
can->Draw();
TH1D *h_side = (TH1D*)can->GetListOfPrimitives()->At(i_h_side[iMethod]);
TH1D *h_signal = (TH1D*)can->GetListOfPrimitives()->At(i_h_signal[iMethod]);
// Simulate different cross section by scaling a certain background
if (iMethod==1) {
TH1D *h_side_temp_scaled = (TH1D*)h_side->Clone(); h_side_temp_scaled->Scale(scale_factors[iSample]);
TH1D *h_signal_temp_scaled = (TH1D*)h_signal->Clone(); h_signal_temp_scaled->Scale(scale_factors[iSample]);
if (iSample==0) {
h_side_sum = h_side_temp_scaled;
h_signal_sum = h_signal_temp_scaled;
} else {
h_side_sum->Add(h_side_temp_scaled);
h_signal_sum->Add(h_signal_temp_scaled);
}
} else if (iMethod==2) {
h_side = h_side_sum;
h_signal = h_signal_sum;
}
TH1D *h_pred =(TH1D*)h_side->Clone();
if (iMethod!=0&&rebin>1) { h_side->Rebin(rebin); h_signal->Rebin(rebin); h_pred->Rebin(rebin); }
TLegend *leg = (TLegend*)can->GetListOfPrimitives()->At(can->GetListOfPrimitives()->GetEntries()-1);
leg->SetX1(0.35); leg->SetX2(0.65); leg->SetY1(0.6);
// Add ratio plot
int y1 = 350;
int y2 = 150;
int mid2 = 10;
can->Divide(1,2);
// Pad 1 (80+500+20 x 40+500)
TVirtualPad* p = can->cd(1);
p->SetPad(0,(y2+60+mid2)/(y1+y2+100.0+mid2),1,1);
p->SetTopMargin(40.0/(y1+40));
p->SetBottomMargin(0);
p->SetRightMargin(0.05);
p->SetLogy(1);
h_side->GetYaxis()->SetRangeUser(1.00001e-4,1e4);
h_side->Draw("HIST");
h_signal->Draw("SAMEHISTE1");
leg->Draw();
// Pad 2 (80+500+20 x 200+60)
p = can->cd(2);
p->SetGrid(0,1);
p->SetPad(0,0,1,(y2+60+mid2)/(y1+y2+100.0+mid2));
p->SetTopMargin(((float)mid2)/(y2+60+mid2));
p->SetBottomMargin(60.0/(y2+60+mid2));
p->SetRightMargin(0.05);
TH1D* ratio = (TH1D*)h_signal->Clone();
TH1D* div = (TH1D*)h_side->Clone();
//ratio->Scale(1/ratio->GetSumOfWeights());
double sum_bins_ratio = iMethod==0 ? ratio->Integral():
ratio->Integral(ratio->FindBin(R_CUT_LOW),ratio->FindBin(Rranges_ABCD[iMethod][3]));
ratio->Scale(1/sum_bins_ratio);
ratio->SetTitleSize(32.0/(y2+60+mid2),"xyz");
ratio->SetLabelSize(20.0/(y2+60+mid2),"xyz");
//ratio->Scale(1/div->GetSumOfWeights());
double sum_bins_div = iMethod==0 ? div->Integral():
div->Integral(div->FindBin(R_CUT_LOW),div->FindBin(Rranges_ABCD[iMethod][3]));
div->Scale(1/sum_bins_div);
ratio->Divide(div);
//ratio->GetYaxis()->SetRangeUser(0,2);
ratio->GetXaxis()->SetTitleOffset(0.7);
ratio->GetYaxis()->SetNdivisions(305);
ratio->GetYaxis()->SetTitle("Ratio (Norm.)");
ratio->GetYaxis()->SetTitleOffset(0.4);
ratio->SetTitleSize(24.0/(y2+60+mid2),"y");
ratio->SetTitle("");
ratio->SetMarkerStyle(20);
ratio->SetMarkerColor(1);
ratio->SetLineColor(1);
ratio->GetYaxis()->SetRangeUser(0,4);
ratio->Draw("PE1");
TLine* l = new TLine(ratio->GetXaxis()->GetXmin(), 1, ratio->GetXaxis()->GetXmax(), 1);
l->SetLineWidth(2);
//l->SetLineColor(2);
l->SetLineStyle(2);
l->Draw();
gPad->Update();
// Fit ratio
//TF1 *fit_ratio;
//if (iMethod==1) {
// fit_ratio = new TF1("fit_ratio","pol0", Rranges_ABCD[iMethod][0], Rranges_ABCD[iMethod][1]);
// fit_ratio->SetLineColor(1);
// ratio->Fit("fit_ratio","RE");
// fit_ratio->SetRange(Rranges_ABCD[iMethod][0], Rranges_ABCD[iMethod][2]);
// fit_ratio->Draw("SAME");
//}
p = can->cd(1);
// calculate integrals
double integral[2][2] = { { 0, 0 }, { 0, 0 } };
double integral_error[2][2] = { { 0, 0 }, { 0, 0 } };
double nevt[2][2] = { { 0, 0 }, { 0, 0 } };
//std::cout<<samples[iMethod][iSample]<<":"<<std::endl;
for (int i=0; i<2; ++i) {
for (int bin=1; bin<=h_side->GetNbinsX(); ++bin) {
if (h_signal->GetXaxis()->GetBinLowEdge(bin)>=Rranges_ABCD[iMethod][i*2] &&
h_signal->GetXaxis()->GetBinUpEdge(bin)<=Rranges_ABCD[iMethod][i*2+1]) {
//std::cout<<bin<<"="<<h_side->GetBinCenter(bin);
//if (i==0) std::cout<<" in, ";
//else std::cout<<" out, ";
double c0 = h_side->GetBinContent(bin), c1 = h_signal->GetBinContent(bin);
double e0 = h_side->GetBinError(bin), e1 = h_signal->GetBinError(bin);
//std::cout<<h_signal->GetBinError(bin)<<" "<<sqrt(c1*weight[iSample])<<std::endl;
if (baderror&&iMethod==1) {
e0 = sqrt(c0*weight[iSample]);
e1 = sqrt(c1*weight[iSample]);
}
nevt[0][i] += (int)(c0*c0/(e0*e0) + 0.5);
nevt[1][i] += (int)(c1*c1/(e1*e1) + 0.5);
integral[0][i] += c0;
integral[1][i] += c1;
//if (iMethod==1) { // weight bin by projected ratio (correction)
// double bincent = h_signal->GetXaxis()->GetBinLowEdge(bin);
// integral[0][1] *= fit_ratio->Eval(bincent);
//}
integral_error[0][i] += e0*e0;
integral_error[1][i] += e1*e1;
//if (iSample==0&&e1>0) std::cout<<bin<<" "<<c1<<" +- "<<e1*e1<<" nevt = "<<((int)(c1*c1/(e1*e1) + 0.5))<<std::endl;
}
}
//if (iSample==1&&i==1) std::cout<<integral[1][i]<<" +- "<<integral_error[1][i]<<std::endl;
integral_error[0][i] = sqrt(integral_error[0][i]);
integral_error[1][i] = sqrt(integral_error[1][i]);
}
//std::cout<<nevt[1][1]<<std::endl;
// predict yields using 2 methods (ABCD and constrained R-shape fit method combined)
// ABCD method
double a = integral[0][0], b = integral[0][1], c = integral[1][0], d = integral[1][1];
double a_err = integral_error[0][0], b_err = integral_error[0][1], c_err = integral_error[1][0], d_err = integral_error[1][1];
// Calculate error
// z = x / y -> z_err = sqrt( (x*x*y_err*y_err + y*y*x_err*x_err)/(y*y*y*y) )
// z = x * y -> z_err = sqrt ( x*x*y_err*y_err + y*y*x_err*x_err )
double c_per_a_err = sqrt((c*c*a_err*a_err + a*a*c_err*c_err)/(a*a*a*a));
double d_abcd = b * (c/a), d_abcd_err = sqrt(b*b*c_per_a_err*c_per_a_err + (c/a)*(c/a)*b_err*b_err);
double d_nevt = nevt[1][1];
double d_ratio = d_abcd / d;
double d_ratio_err = sqrt((d_err/d)*(d_err/d) + (d_abcd_err/d_abcd)*(d_abcd_err/d_abcd))*d_ratio;
double d_pull = (d_abcd-d)/sqrt(d_abcd_err*d_abcd_err + d_err*d_err);
// Scaled plot
h_pred->Scale(c/a);
h_pred->SetLineColor(1);
h_pred->SetLineStyle(2);
h_pred->Draw("SAMEHIST");
leg->AddEntry(h_pred, "Prediction (ABCD)", "l");
double fit_integral[2][2], fit_integral_error[2][2], d_fit_comb = 0, d_fit_comb_err = 0;
if (iMethod==1) {
// Fit in the full range of NTop Sideband
// Do fitting and calculate integrals
TF1 *fit_side = new TF1("NTopSide_fit","exp([0]+[1]*x)", iMethod==2 ? 0.2 : sideband_fit_low_range[iSample], Rranges_ABCD[iMethod][3]);
fit_side->SetLineColor(h_side->GetLineColor());
h_side->Fit("NTopSide_fit","QRE");
//fit_side->Draw("SAME");
double Rranges_ACfit[3] = { sideband_fit_low_range[iSample], Rranges_ABCD[iMethod][2], Rranges_ABCD[iMethod][3] };
for (int i=0; i<2; ++i) {
fit_integral[0][i] = fit_side->Integral(Rranges_ACfit[i], Rranges_ACfit[i+1])/h_signal->GetXaxis()->GetBinWidth(1);
fit_integral_error[0][i] = fit_side->IntegralError(Rranges_ACfit[i], Rranges_ACfit[i+1])/h_signal->GetXaxis()->GetBinWidth(1);
}
double par0 = fit_side->GetParameter(0), par0_error = fit_side->GetParError(0);
double par1 = fit_side->GetParameter(1), par1_error = fit_side->GetParError(1);
double par1min, par1max; fit_side->GetParLimits(1, par1min, par1max);
// Fit in the Signal region
// Fitting in sideband, get B area under curve and scale by C/A
TF1 *fit_signal = new TF1("NTopSignal_RSide_fit","exp([0]+[1]*x)", Rranges_ACfit[0], Rranges_ABCD[iMethod][3]);
fit_signal->SetLineColor(h_signal->GetLineColor());
//fit_signal->SetParameter(1, par1);
//fit_signal->SetParLimits(1, par1min, par1max);
h_signal->Fit("NTopSignal_RSide_fit","QREB");
//fit_signal->Draw("SAME");
for (int i=0; i<2; ++i) {
fit_integral[1][i] = fit_signal->Integral(Rranges_ACfit[i], Rranges_ACfit[i+1])/h_signal->GetXaxis()->GetBinWidth(1);
fit_integral_error[1][i] = fit_signal->IntegralError(Rranges_ACfit[i], Rranges_ACfit[i+1])/h_signal->GetXaxis()->GetBinWidth(1);
}
d_fit_comb = fit_integral[0][1] * (c/a);
d_fit_comb_err = sqrt(fit_integral[0][1]*fit_integral[0][1]*c_per_a_err*c_per_a_err + (c/a)*(c/a)*fit_integral_error[0][1]*fit_integral_error[0][1]);
TF1 *fit_pred = new TF1("Predicted_fit","exp([0]+[1]*x)", Rranges_ACfit[0], Rranges_ACfit[2]);
fit_pred->SetLineColor(1);
fit_pred->SetLineStyle(2);
fit_pred->FixParameter(0, par0+std::log(c/a));
fit_pred->FixParameter(1, par1);
h_signal->Fit("Predicted_fit","QREB+");
//fit_pred->Draw("SAME");
}
// Save plot
if (iMethod==3) samples[iMethod][iSample] = "T1tttt";
std::string name = samples[iMethod][iSample];
if (iMethod==2) name = "AllBkg";
if (save) can->SaveAs((out+"BkgEst/ABCD_closure_"+name+"."+ext).c_str());
// Check compatibility of prediction to observed distribution
double ks_test = h_pred->KolmogorovTest(h_signal);
if (iMethod==1) {
sum_a += a; sum_b += b; sum_c += c; sum_d += d;
sum_a_err += a_err*a_err; sum_b_err += b_err*b_err; sum_c_err += c_err*c_err; sum_d_err += d_err*d_err;
sum_d_abcd += d_abcd; sum_d_abcd_err += d_abcd_err*d_abcd_err;
sum_b_fit += fit_integral[0][1]; sum_b_fit_err += fit_integral_error[0][1]*fit_integral_error[0][1];
sum_d_fit += fit_integral[1][1]; sum_d_fit_err += fit_integral_error[1][1]*fit_integral_error[1][1];
sum_d_fit_comb += d_fit_comb; sum_d_fit_comb_err += d_fit_comb_err*d_fit_comb_err;
sum_d_nevt += d_nevt;
//printf(" %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f |\n", d_fit_comb, d_fit_comb_err, d_abcd, d_abcd_err, d, d_err, d_ratio, d_ratio_err);
}
if (latex) {
printf("%s & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ &", samples[iMethod][iSample].c_str(), a, a_err, b, b_err, c, c_err);
printf(" $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & %.2f & %.2f \\\\\n", d_abcd, d_abcd_err, d, d_err, d_ratio, d_ratio_err, d_pull, ks_test);
} else {
printf("| %s | %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f |", samples[iMethod][iSample].c_str(), a, a_err, b, b_err, c, c_err);
printf(" %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f | %.2f | %.2f |\n", d_abcd, d_abcd_err, d, d_err, d_ratio, d_ratio_err, d_pull, ks_test);
}
// Combining best methods
if ((iMethod==0&&iSample==0)||(iMethod==1&&iSample!=0)) {
comb_d_abcd += d_abcd; comb_d += d; comb_d_abcd_err += d_abcd_err*d_abcd_err; comb_d_err += d_err*d_err;
}
}
if (iMethod==1) {
sum_a_err = sqrt(sum_a_err); sum_b_err = sqrt(sum_b_err); sum_c_err = sqrt(sum_c_err); sum_d_err = sqrt(sum_d_err);
sum_b_fit_err = sqrt(sum_b_fit_err); sum_d_fit_err = sqrt(sum_d_fit_err); sum_d_fit_comb_err = sqrt(sum_d_fit_comb_err);
double sum_d_ratio = sum_d_abcd / sum_d;
double sum_d_ratio_err = sqrt((sum_d_err/sum_d)*(sum_d_err/sum_d) + (sum_d_abcd_err/sum_d_abcd)*(sum_d_abcd_err/sum_d_abcd))*sum_d_ratio;
double sum_d_pull = (sum_d_abcd-sum_d)/sqrt(sum_d_abcd_err*sum_d_abcd_err + sum_d_err*sum_d_err);
if (latex) {
printf("\\hline\n");
printf("Sum Bkg. & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ &", sum_a, sum_a_err, sum_b, sum_b_err, sum_c, sum_c_err);
printf(" $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & %.2f & \\\\\n", sum_d_abcd, sum_d_abcd_err, sum_d, sum_d_err, sum_d_ratio, sum_d_ratio_err, sum_d_pull);
} else {
//printf("| Sum Bkg.| %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f |", sum_a, sum_a_err, sum_b_fit, sum_b_fit_err, sum_b, sum_b_err, sum_c, sum_c_err);
printf("| Sum Bkg.| %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f |", sum_a, sum_a_err, sum_b, sum_b_err, sum_c, sum_c_err);
//printf(" %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f |\n", sum_d_fit_comb, sum_d_fit_comb_err, sum_d_abcd, sum_d_abcd_err, sum_d, sum_d_err, sum_d_ratio, sum_d_ratio_err);
printf(" %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f | %.2f | - |\n", sum_d_abcd, sum_d_abcd_err, sum_d, sum_d_err, sum_d_ratio, sum_d_ratio_err, sum_d_pull);
}
} else if (iMethod==2&&samples[0].size()) {
double comb_d_ratio = comb_d_abcd / comb_d;
double comb_d_ratio_err = sqrt((comb_d_err/comb_d)*(comb_d_err/comb_d) + (comb_d_abcd_err/comb_d_abcd)*(comb_d_abcd_err/comb_d_abcd))*comb_d_ratio;
double comb_d_pull = (comb_d_abcd-comb_d)/sqrt(comb_d_abcd_err*comb_d_abcd_err + comb_d_err*comb_d_err);
if (latex) {
printf("\\hline\n");
printf("\\hline\n");
printf("Combined Bkg. & & & & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & $%.2f \\pm %.2f$ & %.2f & \\\\\n", comb_d_abcd, comb_d_abcd_err, comb_d, comb_d_err, comb_d_ratio, comb_d_ratio_err, comb_d_pull);
printf("\\hline\n");
} else {
printf("| Combined Bkg.| | | | %.2f +- %.2f | %.2f +- %.2f | %.2f +- %.2f | %.2f | - |\n", comb_d_abcd, comb_d_abcd_err, comb_d, comb_d_err, comb_d_ratio, comb_d_ratio_err, comb_d_pull);
}
}
}
if (latex) {
printf("\\hline\n");
printf("\\end{tabular}\n");
printf("\\end{center}\n");
printf("\\end{table*}\n");
}
if (save) gApplication->Terminate();
}