| 1 |
/****
|
| 2 |
|
| 3 |
Off peak status (RestrictToMassPeak) :
|
| 4 |
|
| 5 |
x Necessary adaptations identified
|
| 6 |
x Started working on necessary adaptations
|
| 7 |
x Necessary adaptations implemented
|
| 8 |
x Necessary adaptations tested
|
| 9 |
|
| 10 |
NO ADAPTATIONS REQUIRED
|
| 11 |
|
| 12 |
|
| 13 |
****/
|
| 14 |
#include <iostream>
|
| 15 |
#include <sstream>
|
| 16 |
#include <iomanip>
|
| 17 |
#include <TFile.h>
|
| 18 |
#include <TTree.h>
|
| 19 |
#include <TH1.h>
|
| 20 |
#include <TF1.h>
|
| 21 |
#include <TMath.h>
|
| 22 |
#include <TCanvas.h>
|
| 23 |
#include <vector>
|
| 24 |
#include <TROOT.h>
|
| 25 |
#include <TLine.h>
|
| 26 |
#include <TLegend.h>
|
| 27 |
#include <TLatex.h>
|
| 28 |
#include <TRandom.h>
|
| 29 |
#ifndef GeneralToolBoxLoaded
|
| 30 |
#include "GeneralToolBox.C"
|
| 31 |
#endif
|
| 32 |
#ifndef Verbosity
|
| 33 |
#define Verbosity 0
|
| 34 |
#endif
|
| 35 |
|
| 36 |
using namespace std;
|
| 37 |
|
| 38 |
Double_t LogParabola(Double_t *x,Double_t *par)
|
| 39 |
{
|
| 40 |
return par[0]*TMath::Exp(-par[1]*(x[0]-par[2])*(x[0]-par[2])); // we're adding a "logarithmic parabola" :-)
|
| 41 |
//note: the abs() around the first parameter ensures that, when fitting, no negative values are chosen.
|
| 42 |
}
|
| 43 |
|
| 44 |
Double_t LogParabolaP(Double_t *x,Double_t *par)
|
| 45 |
{
|
| 46 |
float fitval = par[0]*TMath::Exp(-par[1]*(x[0]-par[2])*(x[0]-par[2])); // we're adding a "logarithmic parabola" :-)
|
| 47 |
fitval+= statErrorP(fitval);
|
| 48 |
return fitval;
|
| 49 |
}
|
| 50 |
|
| 51 |
Double_t LogParabolaN(Double_t *x,Double_t *par)
|
| 52 |
{
|
| 53 |
float fitval = par[0]*TMath::Exp(-par[1]*(x[0]-par[2])*(x[0]-par[2])); // we're adding a "logarithmic parabola" :-)
|
| 54 |
fitval-= statErrorN(fitval);
|
| 55 |
return fitval;
|
| 56 |
}
|
| 57 |
|
| 58 |
|
| 59 |
|
| 60 |
bool doreject=false;
|
| 61 |
float low_reject=-10;
|
| 62 |
float hi_reject=10;
|
| 63 |
|
| 64 |
bool dofixed=true;
|
| 65 |
|
| 66 |
|
| 67 |
bool addparabola=true;
|
| 68 |
float parabola_height=0;
|
| 69 |
float parabola_inclination=0;
|
| 70 |
float parabola_pointzero=0;
|
| 71 |
|
| 72 |
float find_KM_peak(TH1F *all, TH1F *ttbar, float minfit, float maxfit, bool is_data, float &error,float &Sigma);
|
| 73 |
float find_Gauss_peak(TH1F *all, TH1F *ttbar, float minfit, float maxfit, bool is_data, float &error,float &Sigma,int numsig);
|
| 74 |
|
| 75 |
Double_t CrystalBallPlusLogParabola(double *x, double *par)
|
| 76 |
{
|
| 77 |
//parameters:
|
| 78 |
//N: the way we scale the function
|
| 79 |
//alpha (where the function changes)
|
| 80 |
//n: exponent of the power expression in the left area
|
| 81 |
//xbar: peak of the gaussian part (RHS)
|
| 82 |
//sigma: width of the gaussian part (RHS)
|
| 83 |
float N=par[0];
|
| 84 |
float alpha=par[3]; //verified (orig: 1)
|
| 85 |
float n=par[4]; // verified (orig: 2)
|
| 86 |
float xbar=par[1]; //verified (orig: 3)
|
| 87 |
float sigma=par[2]; //verified (orig: 4)
|
| 88 |
float altX=-x[0];
|
| 89 |
float result=-999;
|
| 90 |
if(((altX-xbar)/sigma>-alpha)){
|
| 91 |
result=N*TMath::Exp(-(altX-xbar)*(altX-xbar)/(2*sigma*sigma));
|
| 92 |
}
|
| 93 |
else
|
| 94 |
{
|
| 95 |
//if we are outside the central (Gaussian) area things become more difficult ...
|
| 96 |
float A=TMath::Power(n/TMath::Abs(alpha),n)*TMath::Exp(-alpha*alpha/2);
|
| 97 |
float B=n/TMath::Abs(alpha) - TMath::Abs(alpha);
|
| 98 |
if((altX-xbar)/sigma<=-alpha) result=N*A*TMath::Power((B-((altX-xbar)/sigma)),-n);
|
| 99 |
if((altX-xbar)/sigma>=alpha) result=N*A*TMath::Power((B+((altX-xbar)/sigma)),-n);
|
| 100 |
}
|
| 101 |
|
| 102 |
result+=par[5]*TMath::Exp(-par[6]*(x[0]-par[7])*(x[0]-par[7])); // we're adding a "logarithmic parabola" :-)
|
| 103 |
if(par[5]<0) return -999; // there can be no negative ttbar contribution, so just return a value which is going to be a horrible fit.
|
| 104 |
if(par[6]<0) return -999; // the parabola needs to close (i.e. tend to negative values for large |jzb|, not to large positive values)
|
| 105 |
|
| 106 |
return result;
|
| 107 |
}
|
| 108 |
|
| 109 |
Double_t CrystalBallPlusLogParabolaP(double *x, double *par)
|
| 110 |
{
|
| 111 |
float parameter_bkp=par[5];
|
| 112 |
par[5]=0;
|
| 113 |
float zjetsresult=CrystalBallPlusLogParabola(x,par);
|
| 114 |
par[5]=parameter_bkp;
|
| 115 |
parameter_bkp=par[0];
|
| 116 |
par[0]=0;
|
| 117 |
float ttbarresult=CrystalBallPlusLogParabola(x,par);
|
| 118 |
par[0]=parameter_bkp;
|
| 119 |
return zjetsresult+ttbarresult+TMath::Sqrt(zjetsresult+(1.0/3.0)*ttbarresult);
|
| 120 |
}
|
| 121 |
|
| 122 |
Double_t CrystalBallPlusLogParabolaN(double *x, double *par)
|
| 123 |
{
|
| 124 |
float parameter_bkp=par[5];
|
| 125 |
par[5]=0;
|
| 126 |
float zjetsresult=CrystalBallPlusLogParabola(x,par);
|
| 127 |
par[5]=parameter_bkp;
|
| 128 |
parameter_bkp=par[0];
|
| 129 |
par[0]=0;
|
| 130 |
float ttbarresult=CrystalBallPlusLogParabola(x,par);
|
| 131 |
par[0]=parameter_bkp;
|
| 132 |
return zjetsresult+ttbarresult-TMath::Sqrt(zjetsresult+(1.0/3.0)*ttbarresult);
|
| 133 |
}
|
| 134 |
|
| 135 |
Double_t KrystalMallLogPar(double *x, double *par)
|
| 136 |
{
|
| 137 |
//parameters:
|
| 138 |
//N: the way we scale the function
|
| 139 |
//alpha (where the function changes)
|
| 140 |
//n: exponent of the power expression in the left area
|
| 141 |
//xbar: peak of the gaussian part (RHS)
|
| 142 |
//sigma: width of the gaussian part (RHS)
|
| 143 |
float N=par[0];
|
| 144 |
float alpha=par[1];
|
| 145 |
float n=par[2];
|
| 146 |
float xbar=par[3];
|
| 147 |
float sigma=par[4];
|
| 148 |
float altX=x[0];
|
| 149 |
float result=-999;
|
| 150 |
if(doreject&&x[0]>low_reject&&x[0]<hi_reject)
|
| 151 |
{
|
| 152 |
TF1::RejectPoint();
|
| 153 |
return 0;
|
| 154 |
}
|
| 155 |
if(((altX-xbar)/sigma>-alpha)&&((altX-xbar)/sigma<alpha)){
|
| 156 |
result=N*TMath::Exp(-(altX-xbar)*(altX-xbar)/(2*sigma*sigma));
|
| 157 |
}
|
| 158 |
else
|
| 159 |
{
|
| 160 |
//if we are outside the central (Gaussian) area things become more difficult ...
|
| 161 |
float A=TMath::Power(n/TMath::Abs(alpha),n)*TMath::Exp(-alpha*alpha/2);
|
| 162 |
float B=n/TMath::Abs(alpha) - TMath::Abs(alpha);
|
| 163 |
if((altX-xbar)/sigma<=-alpha) result=N*A*TMath::Power((B-((altX-xbar)/sigma)),-n);
|
| 164 |
if((altX-xbar)/sigma>=alpha) result=N*A*TMath::Power((B+((altX-xbar)/sigma)),-n);
|
| 165 |
}
|
| 166 |
|
| 167 |
if(addparabola) {
|
| 168 |
if(dofixed) {
|
| 169 |
result+=parabola_height*TMath::Exp(-parabola_inclination*(x[0]-parabola_pointzero)*(x[0]-parabola_pointzero)); // we're adding a "logarithmic parabola" :-)
|
| 170 |
}
|
| 171 |
else {
|
| 172 |
result+=par[5]*TMath::Exp(-par[6]*(x[0]-par[7])*(x[0]-par[7])); // we're adding a "logarithmic parabola" :-)
|
| 173 |
if(par[5]<0) return -999; // there can be no negative ttbar contribution, so just return a value which is going to be a horrible fit.
|
| 174 |
if(par[6]<0) return -999; // the parabola needs to close (i.e. tend to negative values for large |jzb|, not to large positive values)
|
| 175 |
}
|
| 176 |
}
|
| 177 |
return result;
|
| 178 |
}
|
| 179 |
|
| 180 |
|
| 181 |
void do_ttbar_fit(TH1F *ttbar,TF1 *logpar, TF1 *KM)
|
| 182 |
{
|
| 183 |
logpar->SetParameters(10,2,3);
|
| 184 |
ttbar->Fit(logpar,"NQ");
|
| 185 |
ttbar->Fit(logpar,"NQ");
|
| 186 |
ttbar->Fit(logpar,"NQ");
|
| 187 |
ttbar->Fit(logpar,"NQ");
|
| 188 |
ttbar->SetStats(0);
|
| 189 |
parabola_height=logpar->GetParameter(0);
|
| 190 |
parabola_inclination=logpar->GetParameter(1);
|
| 191 |
parabola_pointzero=logpar->GetParameter(2);
|
| 192 |
}
|
| 193 |
|
| 194 |
void draw_complete_fit(TH1F *all, TH1F *ttbar, float minfit, float maxfit, bool is_data, TF1 *KM)
|
| 195 |
{
|
| 196 |
TCanvas *fitsummary;
|
| 197 |
if(is_data) {
|
| 198 |
fitsummary= new TCanvas("fitsummary","Fit Summary",1000,500);
|
| 199 |
fitsummary->Divide(2,1);
|
| 200 |
}
|
| 201 |
else {
|
| 202 |
fitsummary= new TCanvas("fitsummary","Fit Summary",1200,400);
|
| 203 |
fitsummary->Divide(3,1);
|
| 204 |
}
|
| 205 |
TF1 *logpar = new TF1("logpar",LogParabola,minfit,maxfit,3);
|
| 206 |
|
| 207 |
logpar->SetParameters(KM->GetParameter(5),KM->GetParameter(6),KM->GetParameter(7));
|
| 208 |
logpar->SetLineColor(kOrange);
|
| 209 |
logpar->SetLineStyle(2);
|
| 210 |
if(!is_data)
|
| 211 |
{
|
| 212 |
ttbar->GetXaxis()->SetTitle("JZB (GeV/c)");
|
| 213 |
ttbar->GetYaxis()->SetTitle("events");
|
| 214 |
ttbar->GetXaxis()->CenterTitle();
|
| 215 |
ttbar->GetYaxis()->CenterTitle();
|
| 216 |
ttbar->SetLineColor(kRed);
|
| 217 |
fitsummary->cd(1);
|
| 218 |
ttbar->Draw();
|
| 219 |
fitsummary->cd(1);
|
| 220 |
logpar->Draw("same");
|
| 221 |
TLegend *leg = new TLegend(0.3,0.25,0.65,0.4);
|
| 222 |
leg->AddEntry(ttbar,"t#bar{t} (mc)","l");
|
| 223 |
leg->AddEntry(logpar,"Fit with Log. Parabola","l");
|
| 224 |
leg->SetLineColor(kWhite);
|
| 225 |
leg->SetFillColor(kWhite);
|
| 226 |
leg->Draw();
|
| 227 |
TText *title1=write_title("t#bar{t} Distribution and Fit");
|
| 228 |
title1->Draw();
|
| 229 |
}
|
| 230 |
fitsummary->cd(2-int(is_data));
|
| 231 |
fitsummary->cd(2-int(is_data))->SetLogy(1);
|
| 232 |
all->GetYaxis()->SetTitle("events");
|
| 233 |
all->GetYaxis()->CenterTitle();
|
| 234 |
all->Draw();
|
| 235 |
ttbar->SetLineColor(kRed);
|
| 236 |
if(!is_data) ttbar->Draw("same");
|
| 237 |
KM->SetLineWidth(1);
|
| 238 |
KM->Draw("same");
|
| 239 |
logpar->SetLineWidth(1);
|
| 240 |
logpar->Draw("same");
|
| 241 |
if(!is_data)ttbar->Draw("same");
|
| 242 |
TLegend *leg2 = new TLegend(0.65,0.65,0.89,0.89);
|
| 243 |
if(is_data) leg2->AddEntry(all,"Data","l");
|
| 244 |
else leg2->AddEntry(all,"Stacked MC","l");
|
| 245 |
leg2->AddEntry(KM,"Fitted KM Function","l");
|
| 246 |
if(!is_data) leg2->AddEntry(ttbar,"t#bar{t} MC","l");
|
| 247 |
leg2->AddEntry(logpar,"t#bar{t} (Fit)","l");
|
| 248 |
leg2->SetFillColor(kWhite);
|
| 249 |
leg2->SetLineColor(kWhite);
|
| 250 |
leg2->Draw();
|
| 251 |
TText *title2=write_title("Distribution and Fits (log.)");
|
| 252 |
title2->Draw();
|
| 253 |
fitsummary->cd(3-is_data);
|
| 254 |
all->Draw();
|
| 255 |
KM->Draw("same");
|
| 256 |
float peaklocation=KM->GetParameter(3);
|
| 257 |
TLine *muline = new TLine(peaklocation,0,peaklocation,all->GetMaximum());
|
| 258 |
muline->SetLineColor(kBlue);
|
| 259 |
muline->SetLineStyle(2);
|
| 260 |
muline->Draw();
|
| 261 |
TLegend *leg = new TLegend(0.75,0.75,0.89,0.89);
|
| 262 |
if(is_data) leg2->AddEntry(all,"Data","l");
|
| 263 |
else leg->AddEntry(all,"Stacked MC","l");
|
| 264 |
leg->AddEntry(KM,"Fitted KM Function","l");
|
| 265 |
stringstream mulinelabel;
|
| 266 |
mulinelabel<<"Peak position at #mu="<<peaklocation;
|
| 267 |
leg->AddEntry(muline,mulinelabel.str().c_str(),"l");
|
| 268 |
leg->SetLineColor(kWhite);
|
| 269 |
leg->SetFillColor(kWhite);
|
| 270 |
leg->Draw();
|
| 271 |
mulinelabel<<"+/-"<<TMath::Abs(KM->GetParError(3));
|
| 272 |
TText *title3=write_title("Distribution and Fits");
|
| 273 |
title3->Draw();
|
| 274 |
TText *titlel=write_title_low(mulinelabel.str().c_str());
|
| 275 |
titlel->Draw();
|
| 276 |
|
| 277 |
stringstream printtop;
|
| 278 |
printtop << "#mu="<<std::setprecision(3)<<KM->GetParameter(3)<<"+/-"<<std::setprecision(3)<<KM->GetParError(3);
|
| 279 |
TLatex *toptext = new TLatex(0,all->GetMaximum()*1.3,printtop.str().c_str());
|
| 280 |
toptext->SetTextAlign(22);
|
| 281 |
// toptext->Draw();
|
| 282 |
|
| 283 |
doreject=false;
|
| 284 |
TF1 *wholefitfunc=(TF1*)KM->Clone();
|
| 285 |
doreject=true;
|
| 286 |
wholefitfunc->SetLineColor(kRed);
|
| 287 |
wholefitfunc->SetLineStyle(2);
|
| 288 |
wholefitfunc->Draw("same");
|
| 289 |
|
| 290 |
fitsummary->cd(2-is_data);
|
| 291 |
wholefitfunc->Draw("same");
|
| 292 |
|
| 293 |
if(is_data) CompleteSave(fitsummary, "fit/Fit_Summary_Data");
|
| 294 |
else CompleteSave(fitsummary,"fit/Fit_Summary_MC");
|
| 295 |
|
| 296 |
}
|
| 297 |
|
| 298 |
float Kostas_algorithm(TH1F *hist, float &error, float &sigma, TF1* fitFunc, float lowlimit, float highlimit,bool is_data)
|
| 299 |
{
|
| 300 |
float mean = hist->GetBinCenter( hist->GetMaximumBin());
|
| 301 |
float rms = hist->GetRMS();
|
| 302 |
mean = hist->GetBinCenter( hist->GetMaximumBin());
|
| 303 |
|
| 304 |
fitFunc->SetParameter(1,mean);
|
| 305 |
|
| 306 |
hist->Fit(fitFunc,"QLL0","",mean-10,mean+10);
|
| 307 |
|
| 308 |
mean=fitFunc->GetParameter(1);
|
| 309 |
rms=fitFunc->GetParameter(2);
|
| 310 |
error=fitFunc->GetParError(1);
|
| 311 |
|
| 312 |
bool printOut = false; // print the peak estimate in the i-th iteration
|
| 313 |
|
| 314 |
// --- perform iterations
|
| 315 |
int numIterations=5;
|
| 316 |
|
| 317 |
if(printOut) dout << " ( ";
|
| 318 |
for(int i=1;i<numIterations+1;i++) //--- modify the number of iterations until peak is stable
|
| 319 |
{
|
| 320 |
hist->Fit(fitFunc,"QLLN","same",mean - lowlimit*rms, mean + highlimit*rms); // fit -2 +1 sigma from previous iteration
|
| 321 |
mean=fitFunc->GetParameter(1);
|
| 322 |
fitFunc->SetRange(mean - lowlimit*rms, mean + highlimit*rms);
|
| 323 |
if(printOut) dout << mean << ",";
|
| 324 |
}
|
| 325 |
if(printOut) dout << " ) ";
|
| 326 |
if(printOut) dout << endl;
|
| 327 |
mean=fitFunc->GetParameter(1);
|
| 328 |
sigma=fitFunc->GetParameter(2);
|
| 329 |
error=1.0*fitFunc->GetParError(1);
|
| 330 |
|
| 331 |
// below this point we're merely doing cosmetics :-)
|
| 332 |
TCanvas *peakfitcanvas = new TCanvas("peakfitcanvas","Fitting Canvas");
|
| 333 |
peakfitcanvas->cd();
|
| 334 |
|
| 335 |
hist->SetMinimum(0);
|
| 336 |
if(is_data) hist->Draw("e1");
|
| 337 |
else hist->Draw("histo");
|
| 338 |
fitFunc->SetLineColor(kBlue);
|
| 339 |
fitFunc->SetLineWidth(1);
|
| 340 |
fitFunc->Draw("same");
|
| 341 |
hist->SetStats(0);
|
| 342 |
TLegend *leg;
|
| 343 |
if(is_data) {
|
| 344 |
leg= make_legend("Fit (Data)");
|
| 345 |
leg->AddEntry(hist,"Data","p");
|
| 346 |
}
|
| 347 |
else {
|
| 348 |
leg= make_legend("Fit (MC)");
|
| 349 |
leg->AddEntry(hist,"MC","l");
|
| 350 |
}
|
| 351 |
|
| 352 |
leg->AddEntry(fitFunc,"Fit","l");
|
| 353 |
leg->SetX1(0.7);
|
| 354 |
leg->SetY1(0.7);
|
| 355 |
leg->Draw();
|
| 356 |
|
| 357 |
TText *ftitle=write_text(0.20,0.86,"Fit results:");
|
| 358 |
ftitle->SetTextSize(0.03);
|
| 359 |
ftitle->SetTextAlign(11);
|
| 360 |
stringstream fitresult;
|
| 361 |
fitresult << "#mu=" << std::setprecision(4) << mean << "+/-" << std::setprecision(4) << error;
|
| 362 |
// TText *title1=write_text(0.20,0.96,fitresult.str().c_str());
|
| 363 |
TText *title1=write_text(0.20,0.82,fitresult.str().c_str());
|
| 364 |
title1->SetTextSize(0.03);
|
| 365 |
title1->SetTextAlign(11);
|
| 366 |
stringstream sigmainfo;
|
| 367 |
sigmainfo << "#sigma=" << std::setprecision(4) << fitFunc->GetParameter(2) << "+/-" << std::setprecision(4) << fitFunc->GetParError(2);
|
| 368 |
// TText *sigmatext=write_text(0.80,0.96,sigmainfo.str().c_str());
|
| 369 |
TText *sigmatext=write_text(0.20,0.78,sigmainfo.str().c_str());
|
| 370 |
sigmatext->SetTextSize(0.03);
|
| 371 |
sigmatext->SetTextAlign(11);
|
| 372 |
|
| 373 |
// TText* toptitle;
|
| 374 |
// if(is_data) toptitle = write_title("Fit Result (data)");
|
| 375 |
// else toptitle = write_title("Fit Result (MC)");
|
| 376 |
// toptitle->Draw();
|
| 377 |
ftitle->Draw();
|
| 378 |
title1->Draw();
|
| 379 |
sigmatext->Draw();
|
| 380 |
if(!is_data) {
|
| 381 |
CompleteSave(peakfitcanvas,"fit/Fit_Summary_MC");
|
| 382 |
PlottingSetup::JZBPeakPositionMC=mean;
|
| 383 |
PlottingSetup::JZBPeakWidthMC=fitFunc->GetParameter(2);
|
| 384 |
} else {
|
| 385 |
CompleteSave(peakfitcanvas,"fit/Fit_Summary_Data");
|
| 386 |
PlottingSetup::JZBPeakPositionData=mean;
|
| 387 |
PlottingSetup::JZBPeakWidthData=fitFunc->GetParameter(2);
|
| 388 |
}
|
| 389 |
delete peakfitcanvas;
|
| 390 |
|
| 391 |
return mean;
|
| 392 |
}
|
| 393 |
|
| 394 |
|
| 395 |
|
| 396 |
float find_peak(TH1F *all, TH1F *ttbar, float minfit, float maxfit, bool is_data, float &error,float &Sigma, int method)
|
| 397 |
{
|
| 398 |
float peak_position=0;
|
| 399 |
if(method==0||method>1) {
|
| 400 |
//looking at a gaus request
|
| 401 |
int numsig=1;
|
| 402 |
if(method>1) numsig=method;
|
| 403 |
peak_position=find_Gauss_peak(all,ttbar,minfit,maxfit,is_data,error,Sigma,numsig);
|
| 404 |
}
|
| 405 |
if(method==1) {
|
| 406 |
//looking at a KM request
|
| 407 |
peak_position=find_KM_peak(all,ttbar,minfit,maxfit,is_data,error,Sigma);
|
| 408 |
}
|
| 409 |
if(method==-99) { // KOSTAS!!
|
| 410 |
TF1 *f1 = new TF1("f1","gaus",-40,40);
|
| 411 |
peak_position=Kostas_algorithm(all,error,Sigma,f1,2.5,2.5,is_data);
|
| 412 |
}
|
| 413 |
return peak_position;
|
| 414 |
}
|
| 415 |
|
| 416 |
|
| 417 |
float get_Gaussian_peak(TH1F *hist, float &error, float &sigma, TF1* fitFunc, float lowlimit, float highlimit,bool is_data,int numsig)
|
| 418 |
{
|
| 419 |
TCanvas *fitcanvas = new TCanvas("fitcanvas","fitcanvas");
|
| 420 |
float mean = hist->GetBinCenter( hist->GetMaximumBin());
|
| 421 |
float rms = hist->GetRMS();
|
| 422 |
|
| 423 |
mean = hist->GetBinCenter( hist->GetMaximumBin());
|
| 424 |
|
| 425 |
fitFunc->SetParameter(1,mean);
|
| 426 |
|
| 427 |
hist->Fit(fitFunc,"QLL0","",mean-10,mean+10);
|
| 428 |
|
| 429 |
mean=fitFunc->GetParameter(1);
|
| 430 |
rms=fitFunc->GetParameter(2);
|
| 431 |
error=fitFunc->GetParError(1);
|
| 432 |
|
| 433 |
bool printOut = false; // print the peak estimate in the i-th iteration
|
| 434 |
|
| 435 |
// --- perform iterations
|
| 436 |
int numIterations=5;
|
| 437 |
|
| 438 |
if(printOut) dout << " ( ";
|
| 439 |
for(int i=1;i<numIterations+1;i++) //--- modify the number of iterations until peak is stable
|
| 440 |
{
|
| 441 |
hist->Fit(fitFunc,"QLLN","same",mean - numsig*rms, mean + numsig*rms); // fit -2 +1 sigma from previous iteration
|
| 442 |
mean=fitFunc->GetParameter(1);
|
| 443 |
fitFunc->SetRange(mean - numsig*rms, mean + numsig*rms);
|
| 444 |
if(printOut) dout << mean << ",";
|
| 445 |
}
|
| 446 |
if(printOut) dout << " ) ";
|
| 447 |
if(printOut) dout << endl;
|
| 448 |
mean=fitFunc->GetParameter(1);
|
| 449 |
sigma=fitFunc->GetParameter(2);
|
| 450 |
error=1.0*fitFunc->GetParError(1);
|
| 451 |
fitcanvas->cd();
|
| 452 |
hist->SetMinimum(0);
|
| 453 |
if(is_data) hist->Draw("e1");
|
| 454 |
else hist->Draw("histo");
|
| 455 |
fitFunc->SetLineColor(kBlue);
|
| 456 |
fitFunc->SetLineWidth(1);
|
| 457 |
fitFunc->Draw("same");
|
| 458 |
hist->SetStats(0);
|
| 459 |
TLegend *leg;
|
| 460 |
if(is_data) {
|
| 461 |
leg= make_legend("Fit (Data)");
|
| 462 |
leg->AddEntry(hist,"Data","p");
|
| 463 |
}
|
| 464 |
else {
|
| 465 |
leg= make_legend("Fit (MC)");
|
| 466 |
leg->AddEntry(hist,"MC","l");
|
| 467 |
}
|
| 468 |
|
| 469 |
leg->AddEntry(fitFunc,"Fit","l");
|
| 470 |
leg->Draw();
|
| 471 |
|
| 472 |
TText *ftitle=write_text(0.20,0.86,"Fit results:");
|
| 473 |
ftitle->SetTextSize(0.03);
|
| 474 |
ftitle->SetTextAlign(11);
|
| 475 |
stringstream fitresult;
|
| 476 |
fitresult << "#mu=" << std::setprecision(4) << mean << "+/-" << std::setprecision(4) << error;
|
| 477 |
// TText *title1=write_text(0.20,0.96,fitresult.str().c_str());
|
| 478 |
TText *title1=write_text(0.20,0.82,fitresult.str().c_str());
|
| 479 |
title1->SetTextSize(0.03);
|
| 480 |
title1->SetTextAlign(11);
|
| 481 |
stringstream sigmainfo;
|
| 482 |
sigmainfo << "#sigma=" << std::setprecision(4) << fitFunc->GetParameter(2) << "+/-" << std::setprecision(4) << fitFunc->GetParError(2);
|
| 483 |
// TText *sigmatext=write_text(0.80,0.96,sigmainfo.str().c_str());
|
| 484 |
TText *sigmatext=write_text(0.20,0.78,sigmainfo.str().c_str());
|
| 485 |
sigmatext->SetTextSize(0.03);
|
| 486 |
sigmatext->SetTextAlign(11);
|
| 487 |
|
| 488 |
// TText* toptitle;
|
| 489 |
// if(is_data) toptitle = write_title("Fit Result (data)");
|
| 490 |
// else toptitle = write_title("Fit Result (MC)");
|
| 491 |
// toptitle->Draw();
|
| 492 |
ftitle->Draw();
|
| 493 |
title1->Draw();
|
| 494 |
sigmatext->Draw();
|
| 495 |
if(!is_data) CompleteSave(fitcanvas,"fit/Fit_Summary_MC");
|
| 496 |
if(is_data) CompleteSave(fitcanvas,"fit/Fit_Summary_Data");
|
| 497 |
|
| 498 |
|
| 499 |
// dout << "[" << fitFunc->GetParameter(1) << " , " << fitFunc->GetParError(1) << "]" << endl;
|
| 500 |
return mean;
|
| 501 |
}
|
| 502 |
|
| 503 |
|
| 504 |
float find_Gauss_peak(TH1F *all, TH1F *ttbar, float minfit, float maxfit, bool is_data, float &error,float &Sigma,int numsig)
|
| 505 |
{
|
| 506 |
TF1 *fitfunc = new TF1("fitfunc","gaus",minfit,maxfit);
|
| 507 |
float peakpos = get_Gaussian_peak(all,error,Sigma,fitfunc, minfit, maxfit,is_data,numsig);
|
| 508 |
return peakpos;
|
| 509 |
}
|
| 510 |
|
| 511 |
float find_KM_peak(TH1F *all, TH1F *ttbar, float minfit, float maxfit, bool is_data, float &error,float &Sigma)
|
| 512 |
{
|
| 513 |
all->SetLineColor(kBlue);
|
| 514 |
all->SetStats(0);
|
| 515 |
all->SetTitle("");
|
| 516 |
all->GetXaxis()->SetTitle("JZB (GeV/c)");
|
| 517 |
all->GetYaxis()->SetTitle("events");
|
| 518 |
all->GetXaxis()->CenterTitle();
|
| 519 |
all->GetYaxis()->CenterTitle();
|
| 520 |
TF1 *fitfunc = new TF1("fitfunc",KrystalMallLogPar,0.8*minfit,0.8*maxfit,8);
|
| 521 |
if(!is_data)
|
| 522 |
{
|
| 523 |
TF1 *logpar = new TF1("logpar",LogParabola,minfit,maxfit,3);
|
| 524 |
do_ttbar_fit(ttbar,logpar,fitfunc);
|
| 525 |
fitfunc->SetParameters(1000,2,2.5,-1.6,4,logpar->GetParameter(0),logpar->GetParameter(1),logpar->GetParameter(2));
|
| 526 |
parabola_height=logpar->GetParameter(0);
|
| 527 |
parabola_inclination=logpar->GetParameter(1);
|
| 528 |
parabola_pointzero=logpar->GetParameter(2);
|
| 529 |
dofixed=true;//ttbar is known so we can fix the parameters and don't need to use them for fitting!
|
| 530 |
}
|
| 531 |
else
|
| 532 |
{
|
| 533 |
fitfunc->SetParameters(1000,2,2.5,-1.6,4,5.45039,0.000324593,12.3528);
|
| 534 |
dofixed=false;
|
| 535 |
}
|
| 536 |
|
| 537 |
vector<float> chi2values;
|
| 538 |
addparabola=true;
|
| 539 |
for (int ifit=0;ifit<100;ifit++)
|
| 540 |
{
|
| 541 |
all->Fit(fitfunc,"NQ");
|
| 542 |
chi2values.push_back(fitfunc->GetChisquare());
|
| 543 |
if(ifit>5&&chi2values[ifit-2]==chi2values[ifit]) break;
|
| 544 |
}
|
| 545 |
/*
|
| 546 |
The parameters represent the following quantities:
|
| 547 |
float N=par[0];
|
| 548 |
float alpha=par[1];
|
| 549 |
float n=par[2];
|
| 550 |
float xbar=par[3];
|
| 551 |
float sigma=par[4];
|
| 552 |
*/
|
| 553 |
//we are clearing an area of two sigma to the left and to the right of the center of the function for the "real fit".
|
| 554 |
low_reject=-2*fitfunc->GetParameter(4)+fitfunc->GetParameter(3);
|
| 555 |
hi_reject=fitfunc->GetParameter(3)+2*fitfunc->GetParameter(4);
|
| 556 |
if(low_reject>-15) low_reject=-10;
|
| 557 |
if(hi_reject<15) hi_reject=10;
|
| 558 |
doreject=true;//activating the rejection :-)
|
| 559 |
|
| 560 |
for (int ifit=0;ifit<100;ifit++)
|
| 561 |
{
|
| 562 |
all->Fit(fitfunc,"NQ");
|
| 563 |
chi2values.push_back(fitfunc->GetChisquare());
|
| 564 |
if(ifit>5&&chi2values[ifit-2]==chi2values[ifit]) break;
|
| 565 |
}
|
| 566 |
|
| 567 |
draw_complete_fit(all,ttbar,minfit,maxfit,is_data,fitfunc);
|
| 568 |
doreject=true;
|
| 569 |
error=fitfunc->GetParError(3);
|
| 570 |
Sigma=fitfunc->GetParameter(4);//sigma
|
| 571 |
|
| 572 |
return fitfunc->GetParameter(3);
|
| 573 |
}
|
| 574 |
|
| 575 |
Double_t InvCrystalBall(Double_t *x,Double_t *par)
|
| 576 |
{
|
| 577 |
Double_t arg1=0,arg2=0,A=0,B=0;
|
| 578 |
Double_t f1=0;
|
| 579 |
Double_t f2=0;
|
| 580 |
Double_t lim=0;
|
| 581 |
Double_t fitval=0;
|
| 582 |
Double_t N=0;
|
| 583 |
Double_t n=par[4];
|
| 584 |
|
| 585 |
Double_t invX = -x[0];
|
| 586 |
|
| 587 |
if (par[2] != 0)
|
| 588 |
arg1 = (invX-par[1])/par[2];
|
| 589 |
|
| 590 |
arg2 = ( -pow( TMath::Abs(par[3]) , 2 ) ) / 2 ;
|
| 591 |
|
| 592 |
if (par[3] != 0)
|
| 593 |
A = pow( ( n / TMath::Abs( par[3] ) ) , n) * TMath::Exp(arg2);
|
| 594 |
|
| 595 |
if (par[3] != 0)
|
| 596 |
B = n / TMath::Abs(par[3]) - TMath::Abs(par[3]);
|
| 597 |
|
| 598 |
f1 = TMath::Exp(-0.5*arg1*arg1);
|
| 599 |
if (par[2] != 0)
|
| 600 |
f2 = A * pow( ( B - (invX - par[1])/par[2] ) , -n );
|
| 601 |
|
| 602 |
if (par[2] != 0)
|
| 603 |
lim = ( par[1] - invX ) / par[2] ;
|
| 604 |
|
| 605 |
N = par[0];
|
| 606 |
|
| 607 |
|
| 608 |
|
| 609 |
if(lim < par[3])
|
| 610 |
fitval = N * f1;
|
| 611 |
if(lim >= par[3])
|
| 612 |
fitval = N * f2;
|
| 613 |
|
| 614 |
return fitval;
|
| 615 |
}
|
| 616 |
|
| 617 |
|
| 618 |
Double_t DoubleInvCrystalBall(Double_t *x,Double_t *par)
|
| 619 |
{
|
| 620 |
Double_t arg1=0,arg2=0,A=0,B=0;
|
| 621 |
Double_t f1=0;
|
| 622 |
Double_t f2=0;
|
| 623 |
Double_t lim=0;
|
| 624 |
Double_t fitval=0;
|
| 625 |
Double_t N=0;
|
| 626 |
Double_t n=par[4];
|
| 627 |
|
| 628 |
Double_t Sarg1=0,Sarg2=0,SA=0,SB=0;
|
| 629 |
Double_t Sf1=0;
|
| 630 |
Double_t Sf2=0;
|
| 631 |
Double_t Slim=0;
|
| 632 |
Double_t Sfitval=0;
|
| 633 |
Double_t SN=0;
|
| 634 |
Double_t Sn=par[9];
|
| 635 |
|
| 636 |
Double_t invX = -x[0];
|
| 637 |
|
| 638 |
if (par[2] != 0) arg1 = (invX-par[1])/par[2];
|
| 639 |
|
| 640 |
if (par[7] != 0) Sarg1 = (invX-par[6])/par[7];
|
| 641 |
|
| 642 |
arg2 = ( -pow( TMath::Abs(par[3]) , 2 ) ) / 2 ;
|
| 643 |
Sarg2 = ( -pow( TMath::Abs(par[8]) , 2 ) ) / 2 ;
|
| 644 |
|
| 645 |
if (par[3] != 0) A = pow( ( n / TMath::Abs( par[3] ) ) , n) * TMath::Exp(arg2);
|
| 646 |
if (par[8] != 0) SA = pow( ( Sn / TMath::Abs( par[8] ) ) , Sn) * TMath::Exp(Sarg2);
|
| 647 |
|
| 648 |
if (par[3] != 0) B = n / TMath::Abs(par[3]) - TMath::Abs(par[3]);
|
| 649 |
if (par[8] != 0) SB = Sn / TMath::Abs(par[8]) - TMath::Abs(par[8]);
|
| 650 |
|
| 651 |
f1 = TMath::Exp(-0.5*arg1*arg1);
|
| 652 |
Sf1 = TMath::Exp(-0.5*Sarg1*Sarg1);
|
| 653 |
|
| 654 |
if (par[2] != 0) f2 = A * pow( ( B - (invX - par[1])/par[2] ) , -n );
|
| 655 |
if (par[7] != 0) Sf2 = SA * pow( ( SB - (invX - par[6])/par[7] ) , -Sn );
|
| 656 |
|
| 657 |
if (par[2] != 0) lim = ( par[1] - invX ) / par[2] ;
|
| 658 |
if (par[7] != 0) Slim = ( par[6] - invX ) / par[7] ;
|
| 659 |
|
| 660 |
N = par[0];
|
| 661 |
SN = par[5];
|
| 662 |
|
| 663 |
|
| 664 |
|
| 665 |
if(lim < par[3]) fitval = N * f1;
|
| 666 |
if(lim >= par[3]) fitval = N * f2;
|
| 667 |
|
| 668 |
if(Slim < par[8]) Sfitval = SN * Sf1;
|
| 669 |
if(Slim >= par[8]) Sfitval = SN * Sf2;
|
| 670 |
|
| 671 |
return fitval+Sfitval;
|
| 672 |
}
|
| 673 |
|
| 674 |
Double_t DoubleInvCrystalBallP(Double_t *x,Double_t *par)
|
| 675 |
{
|
| 676 |
Double_t arg1=0,arg2=0,A=0,B=0;
|
| 677 |
Double_t f1=0;
|
| 678 |
Double_t f2=0;
|
| 679 |
Double_t lim=0;
|
| 680 |
Double_t fitval=0;
|
| 681 |
Double_t N=0;
|
| 682 |
Double_t n=par[4];
|
| 683 |
|
| 684 |
Double_t Sarg1=0,Sarg2=0,SA=0,SB=0;
|
| 685 |
Double_t Sf1=0;
|
| 686 |
Double_t Sf2=0;
|
| 687 |
Double_t Slim=0;
|
| 688 |
Double_t Sfitval=0;
|
| 689 |
Double_t SN=0;
|
| 690 |
Double_t Sn=par[9];
|
| 691 |
|
| 692 |
Double_t invX = -x[0];
|
| 693 |
|
| 694 |
if (par[2] != 0) arg1 = (invX-par[1])/par[2];
|
| 695 |
|
| 696 |
if (par[7] != 0) Sarg1 = (invX-par[6])/par[7];
|
| 697 |
|
| 698 |
arg2 = ( -pow( TMath::Abs(par[3]) , 2 ) ) / 2 ;
|
| 699 |
Sarg2 = ( -pow( TMath::Abs(par[8]) , 2 ) ) / 2 ;
|
| 700 |
|
| 701 |
if (par[3] != 0) A = pow( ( n / TMath::Abs( par[3] ) ) , n) * TMath::Exp(arg2);
|
| 702 |
if (par[8] != 0) SA = pow( ( Sn / TMath::Abs( par[8] ) ) , Sn) * TMath::Exp(Sarg2);
|
| 703 |
|
| 704 |
if (par[3] != 0) B = n / TMath::Abs(par[3]) - TMath::Abs(par[3]);
|
| 705 |
if (par[8] != 0) SB = Sn / TMath::Abs(par[8]) - TMath::Abs(par[8]);
|
| 706 |
|
| 707 |
f1 = TMath::Exp(-0.5*arg1*arg1);
|
| 708 |
Sf1 = TMath::Exp(-0.5*Sarg1*Sarg1);
|
| 709 |
|
| 710 |
if (par[2] != 0) f2 = A * pow( ( B - (invX - par[1])/par[2] ) , -n );
|
| 711 |
if (par[7] != 0) Sf2 = SA * pow( ( SB - (invX - par[6])/par[7] ) , -Sn );
|
| 712 |
|
| 713 |
if (par[2] != 0) lim = ( par[1] - invX ) / par[2] ;
|
| 714 |
if (par[7] != 0) Slim = ( par[6] - invX ) / par[7] ;
|
| 715 |
|
| 716 |
N = par[0];
|
| 717 |
SN = par[5];
|
| 718 |
|
| 719 |
|
| 720 |
|
| 721 |
if(lim < par[3]) fitval = N * f1;
|
| 722 |
if(lim >= par[3]) fitval = N * f2;
|
| 723 |
|
| 724 |
if(Slim < par[8]) Sfitval = SN * Sf1;
|
| 725 |
if(Slim >= par[8]) Sfitval = SN * Sf2;
|
| 726 |
|
| 727 |
fitval+=Sfitval;
|
| 728 |
fitval+=statErrorP(fitval);
|
| 729 |
|
| 730 |
return fitval;
|
| 731 |
}
|
| 732 |
|
| 733 |
Double_t DoubleInvCrystalBallN(Double_t *x,Double_t *par)
|
| 734 |
{
|
| 735 |
Double_t arg1=0,arg2=0,A=0,B=0;
|
| 736 |
Double_t f1=0;
|
| 737 |
Double_t f2=0;
|
| 738 |
Double_t lim=0;
|
| 739 |
Double_t fitval=0;
|
| 740 |
Double_t N=0;
|
| 741 |
Double_t n=par[4];
|
| 742 |
|
| 743 |
Double_t Sarg1=0,Sarg2=0,SA=0,SB=0;
|
| 744 |
Double_t Sf1=0;
|
| 745 |
Double_t Sf2=0;
|
| 746 |
Double_t Slim=0;
|
| 747 |
Double_t Sfitval=0;
|
| 748 |
Double_t SN=0;
|
| 749 |
Double_t Sn=par[9];
|
| 750 |
|
| 751 |
Double_t invX = -x[0];
|
| 752 |
|
| 753 |
if (par[2] != 0) arg1 = (invX-par[1])/par[2];
|
| 754 |
|
| 755 |
if (par[7] != 0) Sarg1 = (invX-par[6])/par[7];
|
| 756 |
|
| 757 |
arg2 = ( -pow( TMath::Abs(par[3]) , 2 ) ) / 2 ;
|
| 758 |
Sarg2 = ( -pow( TMath::Abs(par[8]) , 2 ) ) / 2 ;
|
| 759 |
|
| 760 |
if (par[3] != 0) A = pow( ( n / TMath::Abs( par[3] ) ) , n) * TMath::Exp(arg2);
|
| 761 |
if (par[8] != 0) SA = pow( ( Sn / TMath::Abs( par[8] ) ) , Sn) * TMath::Exp(Sarg2);
|
| 762 |
|
| 763 |
if (par[3] != 0) B = n / TMath::Abs(par[3]) - TMath::Abs(par[3]);
|
| 764 |
if (par[8] != 0) SB = Sn / TMath::Abs(par[8]) - TMath::Abs(par[8]);
|
| 765 |
|
| 766 |
f1 = TMath::Exp(-0.5*arg1*arg1);
|
| 767 |
Sf1 = TMath::Exp(-0.5*Sarg1*Sarg1);
|
| 768 |
|
| 769 |
if (par[2] != 0) f2 = A * pow( ( B - (invX - par[1])/par[2] ) , -n );
|
| 770 |
if (par[7] != 0) Sf2 = SA * pow( ( SB - (invX - par[6])/par[7] ) , -Sn );
|
| 771 |
|
| 772 |
if (par[2] != 0) lim = ( par[1] - invX ) / par[2] ;
|
| 773 |
if (par[7] != 0) Slim = ( par[6] - invX ) / par[7] ;
|
| 774 |
|
| 775 |
N = par[0];
|
| 776 |
SN = par[5];
|
| 777 |
|
| 778 |
|
| 779 |
|
| 780 |
if(lim < par[3]) fitval = N * f1;
|
| 781 |
if(lim >= par[3]) fitval = N * f2;
|
| 782 |
|
| 783 |
if(Slim < par[8]) Sfitval = SN * Sf1;
|
| 784 |
if(Slim >= par[8]) Sfitval = SN * Sf2;
|
| 785 |
|
| 786 |
fitval+=Sfitval;
|
| 787 |
fitval-=statErrorN(fitval);
|
| 788 |
|
| 789 |
return fitval;
|
| 790 |
}
|
| 791 |
|
| 792 |
Double_t InvCrystalBallP(Double_t *x,Double_t *par)
|
| 793 |
{
|
| 794 |
Double_t arg1=0,arg2=0,A=0,B=0;
|
| 795 |
Double_t f1=0;
|
| 796 |
Double_t f2=0;
|
| 797 |
Double_t lim=0;
|
| 798 |
Double_t fitval=0;
|
| 799 |
Double_t N=0;
|
| 800 |
Double_t n=par[4];
|
| 801 |
|
| 802 |
Double_t invX = -x[0];
|
| 803 |
|
| 804 |
if (par[2] != 0)
|
| 805 |
arg1 = (invX-par[1])/par[2];
|
| 806 |
|
| 807 |
arg2 = ( -pow( TMath::Abs(par[3]) , 2 ) ) / 2 ;
|
| 808 |
|
| 809 |
if (par[3] != 0)
|
| 810 |
A = pow( ( n / TMath::Abs( par[3] ) ) , n) * TMath::Exp(arg2);
|
| 811 |
|
| 812 |
if (par[3] != 0)
|
| 813 |
B = n / TMath::Abs(par[3]) - TMath::Abs(par[3]);
|
| 814 |
|
| 815 |
f1 = TMath::Exp(-0.5*arg1*arg1);
|
| 816 |
if (par[2] != 0)
|
| 817 |
f2 = A * pow( ( B - (invX - par[1])/par[2] ) , -n );
|
| 818 |
|
| 819 |
if (par[2] != 0)
|
| 820 |
lim = ( par[1] - invX ) / par[2] ;
|
| 821 |
|
| 822 |
N = par[0];
|
| 823 |
|
| 824 |
|
| 825 |
|
| 826 |
if(lim < par[3])
|
| 827 |
fitval = N * f1;
|
| 828 |
if(lim >= par[3])
|
| 829 |
fitval = N * f2;
|
| 830 |
|
| 831 |
fitval+= statErrorP(fitval);
|
| 832 |
return fitval;
|
| 833 |
}
|
| 834 |
|
| 835 |
Double_t InvCrystalBallN(Double_t *x,Double_t *par)
|
| 836 |
{
|
| 837 |
Double_t arg1=0,arg2=0,A=0,B=0;
|
| 838 |
Double_t f1=0;
|
| 839 |
Double_t f2=0;
|
| 840 |
Double_t lim=0;
|
| 841 |
Double_t fitval=0;
|
| 842 |
Double_t N=0;
|
| 843 |
Double_t n=par[4];
|
| 844 |
|
| 845 |
Double_t invX = -x[0];
|
| 846 |
|
| 847 |
if (par[2] != 0)
|
| 848 |
arg1 = (invX-par[1])/par[2];
|
| 849 |
|
| 850 |
arg2 = ( -pow( TMath::Abs(par[3]) , 2 ) ) / 2 ;
|
| 851 |
|
| 852 |
if (par[3] != 0)
|
| 853 |
A = pow( ( n / TMath::Abs( par[3] ) ) , n) * TMath::Exp(arg2);
|
| 854 |
|
| 855 |
if (par[3] != 0)
|
| 856 |
B = n / TMath::Abs(par[3]) - TMath::Abs(par[3]);
|
| 857 |
|
| 858 |
f1 = TMath::Exp(-0.5*arg1*arg1);
|
| 859 |
if (par[2] != 0)
|
| 860 |
f2 = A * pow( ( B - (invX - par[1])/par[2] ) , -n );
|
| 861 |
|
| 862 |
if (par[2] != 0)
|
| 863 |
lim = ( par[1] - invX ) / par[2] ;
|
| 864 |
|
| 865 |
N = par[0];
|
| 866 |
|
| 867 |
|
| 868 |
|
| 869 |
if(lim < par[3])
|
| 870 |
fitval = N * f1;
|
| 871 |
if(lim >= par[3])
|
| 872 |
fitval = N * f2;
|
| 873 |
|
| 874 |
fitval-= statErrorN(fitval);
|
| 875 |
return fitval;
|
| 876 |
}
|
| 877 |
|
| 878 |
|
