| 1 |
/*
|
| 2 |
|
| 3 |
g++ ShapeFit.C -o ShapeFit.exec `root-config --glibs --cflags` && ./ShapeFit.exec
|
| 4 |
|
| 5 |
*/
|
| 6 |
|
| 7 |
#include <iostream>
|
| 8 |
#include <vector>
|
| 9 |
#include <fstream>
|
| 10 |
|
| 11 |
#include <TH1.h>
|
| 12 |
#include <TCanvas.h>
|
| 13 |
#include <TObjArray.h>
|
| 14 |
#include <TF1.h>
|
| 15 |
#include <TError.h>
|
| 16 |
#include <TFractionFitter.h>
|
| 17 |
#include <TMath.h>
|
| 18 |
#include <TROOT.h>
|
| 19 |
#include <THStack.h>
|
| 20 |
#include <TLine.h>
|
| 21 |
#include <TColor.h>
|
| 22 |
|
| 23 |
#ifndef ValueClassLoaded
|
| 24 |
#include "ValueClass.C"
|
| 25 |
#endif
|
| 26 |
#ifndef GeneralToolBoxLoaded
|
| 27 |
#include "GeneralToolBox.C"
|
| 28 |
#endif
|
| 29 |
using namespace std;
|
| 30 |
|
| 31 |
void check_compatibility_with_zero(Value sigresult, ofstream &file) {
|
| 32 |
dout << "Checking compatibility with zero for f=" << sigresult << endl;
|
| 33 |
dout << " The result is compatible with zero to " << 1-(TMath::Erf((sigresult.getValue()/(TMath::Sqrt(2)*sigresult.getError())))) << endl;
|
| 34 |
file << "Checking compatibility with zero for f=" << sigresult << endl;
|
| 35 |
file << " The result is compatible with zero to " << 1-(TMath::Erf((sigresult.getValue()/(TMath::Sqrt(2)*sigresult.getError())))) << endl;
|
| 36 |
|
| 37 |
}
|
| 38 |
|
| 39 |
/*
|
| 40 |
vector<float> compute_sigmas(TH1 *signal, float weight) {
|
| 41 |
TH1F *histo = (TH1F*)signal->Clone();
|
| 42 |
histo->Scale(weight);
|
| 43 |
float integral[7];
|
| 44 |
for(int i=0;i<4;i++) integral[i]=0;
|
| 45 |
for(int i=0;i<=histo->GetNbinsX()+1;i++) {
|
| 46 |
integral[0]+=histo->GetBinContent(i);
|
| 47 |
integral[1]+=histo->GetBinContent(i)+histo->GetBinError(i);
|
| 48 |
integral[2]+=histo->GetBinContent(i)+2*histo->GetBinError(i);
|
| 49 |
integral[3]+=histo->GetBinContent(i)+3*histo->GetBinError(i);
|
| 50 |
integral[4]+=histo->GetBinContent(i)+4*histo->GetBinError(i);
|
| 51 |
integral[5]+=histo->GetBinContent(i)+5*histo->GetBinError(i);
|
| 52 |
integral[6]+=histo->GetBinContent(i)+6*histo->GetBinError(i);
|
| 53 |
}
|
| 54 |
float meanintegral = signal->Integral();
|
| 55 |
vector<float> results;
|
| 56 |
results.push_back(integral[0]/meanintegral);
|
| 57 |
results.push_back(integral[1]/meanintegral);
|
| 58 |
results.push_back(integral[2]/meanintegral);
|
| 59 |
results.push_back(integral[3]/meanintegral);
|
| 60 |
results.push_back(integral[4]/meanintegral);
|
| 61 |
results.push_back(integral[5]/meanintegral);
|
| 62 |
results.push_back(integral[6]/meanintegral);
|
| 63 |
return results;
|
| 64 |
}
|
| 65 |
*/
|
| 66 |
void present_limit_fitting_results(TH1 *result,TH1 *data, TH1 *background, TH1 *signal, float bgw, float siw,string filename) {
|
| 67 |
TH1F *bg = (TH1F*)background->Clone();
|
| 68 |
bg->Scale(bgw);
|
| 69 |
TH1F *si = (TH1F*)signal->Clone();
|
| 70 |
si->Scale(siw);
|
| 71 |
// TCanvas *c1 = new TCanvas("c1","c1",500,900);
|
| 72 |
TCanvas *c1 = new TCanvas("c1","c1",1500,500);
|
| 73 |
c1->Divide(3,1);
|
| 74 |
// c1->Divide(1,3);
|
| 75 |
|
| 76 |
c1->cd(1);
|
| 77 |
data->SetTitle("");
|
| 78 |
data->SetStats(0);
|
| 79 |
background->SetLineColor(kBlue);
|
| 80 |
data->SetLineColor(kBlack);
|
| 81 |
signal->SetLineColor(kRed);
|
| 82 |
// background->SetFillColor(TColor::GetColor("#5882FA"));
|
| 83 |
background->SetMarkerSize(0);
|
| 84 |
// background->SetFillColor(TColor::GetColor("#5882FA"));
|
| 85 |
// signal->SetFillColor(TColor::GetColor("#81F781"));
|
| 86 |
signal->SetMarkerSize(0);
|
| 87 |
// signal->SetFillColor(TColor::GetColor("#FE2E2E"));
|
| 88 |
TLegend *leg = new TLegend(0.7,0.7,0.89,0.89);
|
| 89 |
leg->AddEntry(data,"Data","p");
|
| 90 |
leg->AddEntry(bg,"Z+Jets","f");
|
| 91 |
leg->AddEntry(si,"signal","f");
|
| 92 |
leg->SetLineColor(kWhite);
|
| 93 |
leg->SetFillColor(kWhite);
|
| 94 |
data->Draw();
|
| 95 |
background->Draw("histo,same");
|
| 96 |
signal->Draw("histo,same");
|
| 97 |
data->Draw("same");
|
| 98 |
leg->Draw("same");
|
| 99 |
TText* title = write_title("Before fitting");
|
| 100 |
title->Draw();
|
| 101 |
|
| 102 |
c1->cd(2);
|
| 103 |
c1->cd(2)->SetLogy(1);
|
| 104 |
TH1F *datac = (TH1F*)data->Clone();
|
| 105 |
bg->SetFillColor(TColor::GetColor("#5882FA"));
|
| 106 |
bg->SetLineColor(kBlue);
|
| 107 |
// si->SetLineColor(kGreen);
|
| 108 |
// si->SetFillColor(TColor::GetColor("#81F781"));
|
| 109 |
si->SetLineColor(kRed);
|
| 110 |
si->SetFillColor(TColor::GetColor("#FE2E2E"));
|
| 111 |
THStack fresult("fresult","");
|
| 112 |
fresult.Add(bg);
|
| 113 |
fresult.Add(si);
|
| 114 |
fresult.Draw("histo");
|
| 115 |
datac->SetMinimum(0.4);
|
| 116 |
datac->SetMaximum(data->GetMaximum()*10);
|
| 117 |
datac->Draw();
|
| 118 |
fresult.Draw("histo,same");
|
| 119 |
datac->Draw("same");
|
| 120 |
leg->Draw("same");
|
| 121 |
TText* title2 = write_title("After fitting (composition)");
|
| 122 |
title2->Draw();
|
| 123 |
|
| 124 |
c1->cd(3);
|
| 125 |
TH1F *ratio = (TH1F*)result->Clone();
|
| 126 |
ratio->Divide(data);
|
| 127 |
ratio->GetYaxis()->SetRangeUser(0,2);
|
| 128 |
TLine *oneline = new TLine(data->GetBinLowEdge(1),1,data->GetBinLowEdge(data->GetNbinsX())+data->GetBinWidth(data->GetNbinsX()),1);
|
| 129 |
oneline->SetLineColor(kBlue);
|
| 130 |
ratio->SetFillColor(TColor::GetColor("#FA8258"));
|
| 131 |
ratio->SetTitle("");
|
| 132 |
ratio->SetStats(0);
|
| 133 |
ratio->GetYaxis()->SetTitle("(Fit result)/data");
|
| 134 |
ratio->GetYaxis()->CenterTitle();
|
| 135 |
ratio->SetMarkerSize(0);
|
| 136 |
ratio->Draw("E5");
|
| 137 |
oneline->Draw();
|
| 138 |
TText* title3 = write_title("Ratio");
|
| 139 |
title3->Draw();
|
| 140 |
CompleteSave(c1,("upper_limit/"+filename).c_str());
|
| 141 |
CompleteSave(c1->cd(1),("upper_limit/"+filename+"_cd1").c_str());
|
| 142 |
CompleteSave(c1->cd(2),("upper_limit/"+filename+"_cd2").c_str());
|
| 143 |
CompleteSave(c1->cd(3),("upper_limit/"+filename+"_cd3").c_str());
|
| 144 |
|
| 145 |
}
|
| 146 |
|
| 147 |
float compute_upper_limit_based_on_fit(Value signalresult, Value originalfactor, int nsigma) {
|
| 148 |
float stretchby=signalresult.getValue()*(signalresult.getValue()+nsigma*signalresult.getError())/(signalresult.getValue());
|
| 149 |
stretchby/=originalfactor.getValue();
|
| 150 |
return stretchby;
|
| 151 |
}
|
| 152 |
|
| 153 |
|
| 154 |
vector<float> establish_upper_limits(TH1 *data, TH1 *background, TH1 *signal,string filename, ofstream &file) {
|
| 155 |
TObjArray *allmc = new TObjArray(2);
|
| 156 |
allmc->Add(background);
|
| 157 |
allmc->Add(signal);
|
| 158 |
|
| 159 |
TCanvas *c1 = new TCanvas("c1","c1");
|
| 160 |
background->SetMarkerColor(kBlue);
|
| 161 |
background->Draw();
|
| 162 |
signal->SetMarkerColor(kGreen);
|
| 163 |
signal->Draw("same");
|
| 164 |
data->Draw("same");
|
| 165 |
c1->SaveAs("Input_for_establish_upper_limits.png");
|
| 166 |
|
| 167 |
TFractionFitter* fit = new TFractionFitter(data, allmc);
|
| 168 |
|
| 169 |
fit->Constrain(1,1.0,1.0);
|
| 170 |
fit->Constrain(2,0.0,1.1);
|
| 171 |
//fit->Constrain(1,0.0,10.0);
|
| 172 |
TFile *rawinput = new TFile("rawtfitterinput.root","RECREATE");
|
| 173 |
data->Write();
|
| 174 |
signal->Write();
|
| 175 |
background->Write();
|
| 176 |
rawinput->Close();
|
| 177 |
|
| 178 |
Int_t status = fit->Fit();
|
| 179 |
vector<float> results;
|
| 180 |
if(status!=0) return results;
|
| 181 |
|
| 182 |
Double_t mean[2],error[2];
|
| 183 |
fit->GetResult(0,mean[0],error[0]);
|
| 184 |
Value SMresult(mean[0],error[0]);
|
| 185 |
fit->GetResult(1,mean[1],error[1]);
|
| 186 |
Value sigresult(mean[1],error[1]);
|
| 187 |
dout << "Parameters: " << mean[0] << " , " << mean[1] << endl;
|
| 188 |
if(mean[0]<0||mean[1]<0) {
|
| 189 |
dout << "Fitting failed, one of the parameters is negative !!!!" << endl;
|
| 190 |
return results;
|
| 191 |
}
|
| 192 |
|
| 193 |
Value vdata(data->Integral(),TMath::Sqrt(data->Integral()));
|
| 194 |
Value vsig(signal->Integral(),TMath::Sqrt(signal->Integral()));
|
| 195 |
Value vbg(background->Integral(),TMath::Sqrt(background->Integral()));
|
| 196 |
float originalfactor=(signal->Integral())/(data->Integral());
|
| 197 |
float originalfactorSM=(background->Integral())/(data->Integral());
|
| 198 |
|
| 199 |
Value ofac = vsig/vdata;
|
| 200 |
Value ofacsm = vbg /vdata;
|
| 201 |
|
| 202 |
Value stretchsm = SMresult/ofacsm;
|
| 203 |
Value stretchsig = sigresult/ofac;
|
| 204 |
dout << "Data consists of: " << endl;
|
| 205 |
dout << " SM : " << SMresult << endl;
|
| 206 |
dout << " signal: " << sigresult << endl;
|
| 207 |
dout << "This means that the components have been scaled by: " << endl;
|
| 208 |
dout << " SM: has been stretched by " << (mean[0])/originalfactorSM << " "<< stretchsm<<endl;
|
| 209 |
dout << " Signal: has been stretched by " << (mean[1])/originalfactor << " "<< stretchsig<<endl;
|
| 210 |
|
| 211 |
file << "------------------------------------------"<<endl;
|
| 212 |
file << "Now considering: Everything pertaining to " << filename << endl;
|
| 213 |
file << "Data consists of: " << endl;
|
| 214 |
file << " SM : " << SMresult << endl;
|
| 215 |
file << " signal: " << sigresult << endl;
|
| 216 |
file << "This means that the components have been scaled by: " << endl;
|
| 217 |
file << " SM: has been stretched by " << stretchsm<<endl;
|
| 218 |
file << " Signal: has been stretched by " << stretchsig<<endl;
|
| 219 |
|
| 220 |
present_limit_fitting_results((TH1F*)fit->GetPlot(),data,background,signal,(mean[0])/originalfactorSM,(mean[1])/originalfactor,filename);
|
| 221 |
|
| 222 |
check_compatibility_with_zero(sigresult,file);
|
| 223 |
//vector<float> results = compute_sigmas(signal,(mean[1])/originalfactor);
|
| 224 |
|
| 225 |
dout << "\\Chi^2 / ndf = " << fit->GetChisquare() << "/" << fit->GetNDF() << endl;
|
| 226 |
|
| 227 |
results.push_back(compute_upper_limit_based_on_fit(sigresult,ofac,0));
|
| 228 |
results.push_back(compute_upper_limit_based_on_fit(sigresult,ofac,1));
|
| 229 |
results.push_back(compute_upper_limit_based_on_fit(sigresult,ofac,2));
|
| 230 |
results.push_back(compute_upper_limit_based_on_fit(sigresult,ofac,3));
|
| 231 |
|
| 232 |
dout << "Limits: (in terms of x prediction) " << endl;
|
| 233 |
file << "Limits: (in terms of x prediction) " << endl;
|
| 234 |
for (int i=0;i<=3;i++) {
|
| 235 |
dout << i << " sigma : " << results[i] << endl;
|
| 236 |
file << i << " sigma : " << results[i] << endl;
|
| 237 |
}
|
| 238 |
|
| 239 |
|
| 240 |
dout << endl << endl;
|
| 241 |
file << endl << endl;
|
| 242 |
|
| 243 |
if(results[3]<1) {
|
| 244 |
dout << "The point has been excluded, since the upper limit corresponding to 3 sigma, " << results[3] << " is < 1" << endl;
|
| 245 |
file << "The point has been excluded, since the upper limit corresponding to 3 sigma, " << results[3] << " is < 1" << endl;
|
| 246 |
//the ANSII code below will make an X for the excluded point (to celebrate)
|
| 247 |
dout << " __ __" << endl;
|
| 248 |
dout << " \\ \\ / /" << endl;
|
| 249 |
dout << " \\ V / " << endl;
|
| 250 |
dout << " / \\ " << endl;
|
| 251 |
dout << " / /^\\ \\" << endl;
|
| 252 |
dout << " \\/ \\/" << endl;
|
| 253 |
|
| 254 |
file << " __ __" << endl;
|
| 255 |
file << " \\ \\ / /" << endl;
|
| 256 |
file << " \\ V / " << endl;
|
| 257 |
file << " / \\ " << endl;
|
| 258 |
file << " / /^\\ \\" << endl;
|
| 259 |
file << " \\/ \\/" << endl;
|
| 260 |
|
| 261 |
|
| 262 |
}
|
| 263 |
else {
|
| 264 |
dout << "The 3 sigma limit on this point is " << results[3] << " x prediction" << endl;
|
| 265 |
file << "The point has been excluded, since " << results[3] << " x prediction" << endl;
|
| 266 |
}
|
| 267 |
|
| 268 |
|
| 269 |
return results;
|
| 270 |
}
|
| 271 |
|
| 272 |
/*
|
| 273 |
|
| 274 |
Please use the function like this :
|
| 275 |
|
| 276 |
int main() {
|
| 277 |
gROOT->SetStyle("Plain");
|
| 278 |
ofstream myfile;
|
| 279 |
myfile.open ("ShapeFit_log.txt");
|
| 280 |
TH1F *data = new TH1F("data","data",10,0,10);
|
| 281 |
data->GetXaxis()->SetTitle("some variable");
|
| 282 |
data->GetXaxis()->CenterTitle();
|
| 283 |
data->GetYaxis()->SetTitle("events");
|
| 284 |
data->GetYaxis()->CenterTitle();
|
| 285 |
data->Sumw2();
|
| 286 |
TH1F *mc = new TH1F("mc","mc",10,0,10);
|
| 287 |
mc->Sumw2();
|
| 288 |
TH1F *si = new TH1F("si","signal",10,0,10);
|
| 289 |
si->Sumw2();
|
| 290 |
TF1 *f1 = new TF1("f1", "[0] +[1]*x +gaus(2)", 0, 5);
|
| 291 |
f1->SetParameters(6, -1,8, 0.1, 0.05);
|
| 292 |
TF1 *f2 = new TF1("f2", "gaus(2)", 0, 5);
|
| 293 |
f2->SetParameters(5, 0.1, 0.05);
|
| 294 |
TF1 *f3 = new TF1("f3", "[0] +[1]*x", 0, 5);
|
| 295 |
f3->SetParameters(6, -1,8);
|
| 296 |
|
| 297 |
data->FillRandom("f1",10000);
|
| 298 |
mc->FillRandom("f3",6000);
|
| 299 |
si->FillRandom("f2",2000);
|
| 300 |
|
| 301 |
establish_upper_limits(data,mc,si,"test2.png",myfile);
|
| 302 |
myfile.close();
|
| 303 |
return 0;
|
| 304 |
}
|
| 305 |
*/
|
