| 1 |
// $Id: TPFuncs.cc,v 1.2 2009/06/11 20:26:57 loizides Exp $
|
| 2 |
|
| 3 |
#include "MitCommon/MathTools/interface/TPFuncs.h"
|
| 4 |
#include <TF1.h>
|
| 5 |
|
| 6 |
using namespace mithep;
|
| 7 |
|
| 8 |
//--------------------------------------------------------------------------------------------------
|
| 9 |
Double_t TPFuncs::BreitGaus(Double_t x, Double_t m, Double_t mwidth, Double_t msig,
|
| 10 |
Double_t fintf, Double_t xmin, Double_t xmax)
|
| 11 |
{
|
| 12 |
// Breit-Wigner convoluted with Gaussian.
|
| 13 |
// Fit parameters:
|
| 14 |
// m: Most probable location) Breit mean
|
| 15 |
// mwidth: Width (scale) Breit-Wigner
|
| 16 |
// msig: Width (sigma) of convoluted Gaussian function
|
| 17 |
// fintf: interference fraction between Z and gamma
|
| 18 |
// xmin: minimum x (for normalization)
|
| 19 |
// xmax: maximum x (for normalization)
|
| 20 |
|
| 21 |
// Numeric constants
|
| 22 |
const Double_t invsq2pi = 0.3989422804014; //(2pi)^(-1/2)
|
| 23 |
|
| 24 |
// Control constants
|
| 25 |
const Double_t np = 300; // number of convolution steps
|
| 26 |
const Double_t sc = 3.0; // convolution extends to +-sc Gaussian sigmas
|
| 27 |
|
| 28 |
// Range of convolution integral
|
| 29 |
const Double_t xlow = x - sc * msig;
|
| 30 |
const Double_t xupp = x + sc * msig;
|
| 31 |
const Double_t step = (xupp-xlow) / np;
|
| 32 |
|
| 33 |
// Convolution integral of Breit and Gaussian by sum
|
| 34 |
Double_t sum0 = 0.0;
|
| 35 |
if (fintf<1) {
|
| 36 |
for(Double_t i=1.0; i<=np/2; i++) {
|
| 37 |
Double_t xx = xlow + (i-0.5) * step;
|
| 38 |
Double_t fland = BreitWignerZ(xx,m,mwidth);
|
| 39 |
sum0 += fland * TMath::Gaus(x,xx,msig);
|
| 40 |
xx = xupp - (i-0.5) * step;
|
| 41 |
fland = BreitWignerZ(xx,m,mwidth);
|
| 42 |
sum0 += fland * TMath::Gaus(x,xx,msig);
|
| 43 |
}
|
| 44 |
sum0 = (step * sum0 * invsq2pi / msig);
|
| 45 |
}
|
| 46 |
|
| 47 |
Double_t sum1 = 0.0;
|
| 48 |
if (fintf>0) {
|
| 49 |
for(Double_t i=1.0; i<=np/2; i++) {
|
| 50 |
Double_t xx = xlow + (i-0.5) * step;
|
| 51 |
Double_t fland = BreitWignerGamma(xx,m,mwidth);
|
| 52 |
sum1 += fland * TMath::Gaus(x,xx,msig);
|
| 53 |
xx = xupp - (i-0.5) * step;
|
| 54 |
fland = BreitWignerGamma(xx,m,mwidth);
|
| 55 |
sum1 += fland * TMath::Gaus(x,xx,msig);
|
| 56 |
}
|
| 57 |
sum1 = (step * sum1 * invsq2pi / msig / (xmax-xmin));
|
| 58 |
}
|
| 59 |
return ((1.-fintf)*sum0+fintf*sum1);
|
| 60 |
}
|
| 61 |
|
| 62 |
//--------------------------------------------------------------------------------------------------
|
| 63 |
Double_t TPFuncs::BreitGaus(Double_t *x, Double_t *par)
|
| 64 |
{
|
| 65 |
// Breit-Wigner convoluted with Gaussian.
|
| 66 |
// Fit parameters:
|
| 67 |
// norm: Normalization factor
|
| 68 |
// m: Most probable location) Breit mean
|
| 69 |
// mwidth: Width (scale) Breit-Wigner
|
| 70 |
// msig: Width (sigma) of convoluted Gaussian function
|
| 71 |
// fintf: interference fraction between Z and gamma
|
| 72 |
// xmin: minimum x (for normalization)
|
| 73 |
// xmax: maximum x (for normalization)
|
| 74 |
|
| 75 |
Double_t xx = x[0];
|
| 76 |
Double_t norm = par[0];
|
| 77 |
Double_t m = par[1];
|
| 78 |
Double_t mwidth = par[2];
|
| 79 |
Double_t msig = par[3];
|
| 80 |
Double_t fintf = par[4];
|
| 81 |
Double_t xmin = par[5];
|
| 82 |
Double_t xmax = par[6];
|
| 83 |
Double_t ret = norm * BreitGaus(xx, m, mwidth, msig, fintf, xmin, xmax);
|
| 84 |
return ret;
|
| 85 |
}
|
| 86 |
|
| 87 |
//--------------------------------------------------------------------------------------------------
|
| 88 |
Double_t TPFuncs::BreitWignerZ(Double_t x, Double_t mean, Double_t gamma)
|
| 89 |
{
|
| 90 |
// Breit-Wigner for Z peak.
|
| 91 |
|
| 92 |
Double_t bw = (gamma*mean*mean)/
|
| 93 |
((x*x-mean*mean)*(x*x-mean*mean) + x*x*x*x*(gamma*gamma)/(mean*mean));
|
| 94 |
return bw / (TMath::Pi()/2.0);
|
| 95 |
}
|
| 96 |
|
| 97 |
//--------------------------------------------------------------------------------------------------
|
| 98 |
Double_t TPFuncs::BreitWignerZ(Double_t *x, Double_t *par)
|
| 99 |
{
|
| 100 |
// Breit-Wigner for Z peak.
|
| 101 |
|
| 102 |
Double_t xx = x[0];
|
| 103 |
Double_t norm = par[0];
|
| 104 |
Double_t mean = par[1];
|
| 105 |
Double_t gamma = par[2];
|
| 106 |
|
| 107 |
Double_t ret = norm * BreitWignerZ(xx,mean,gamma);
|
| 108 |
return ret;
|
| 109 |
}
|
| 110 |
|
| 111 |
//--------------------------------------------------------------------------------------------------
|
| 112 |
Double_t TPFuncs::BreitWignerGamma(Double_t x, Double_t mean, Double_t gamma)
|
| 113 |
{
|
| 114 |
// Breit-Wigner for gamma interference.
|
| 115 |
|
| 116 |
Double_t bw = ((x*x-mean*mean)*(x*x-mean*mean))/
|
| 117 |
((x*x-mean*mean)*(x*x-mean*mean) + x*x*x*x*(gamma*gamma)/(mean*mean));
|
| 118 |
return bw;
|
| 119 |
}
|
| 120 |
|
| 121 |
//--------------------------------------------------------------------------------------------------
|
| 122 |
Double_t TPFuncs::BreitWignerGamma(Double_t *x, Double_t *par)
|
| 123 |
{
|
| 124 |
// Breit-Wigner for gamma interference.
|
| 125 |
|
| 126 |
Double_t xx = x[0];
|
| 127 |
Double_t norm = par[0];
|
| 128 |
Double_t mean = par[1];
|
| 129 |
Double_t gamma = par[2];
|
| 130 |
|
| 131 |
Double_t ret = norm * BreitWignerGamma(xx,mean,gamma);
|
| 132 |
return ret;
|
| 133 |
}
|
| 134 |
|
| 135 |
//--------------------------------------------------------------------------------------------------
|
| 136 |
Double_t TPFuncs::ExpRange(Double_t mass, Double_t lambda, Double_t xmin, Double_t xmax)
|
| 137 |
{
|
| 138 |
// Probability of an exponential in a given interval.
|
| 139 |
|
| 140 |
if (mass < xmin || mass > xmax)
|
| 141 |
return 0.0;
|
| 142 |
|
| 143 |
if (lambda == 0.0)
|
| 144 |
return 1.0/(xmax-xmin);
|
| 145 |
|
| 146 |
Double_t xmed = 0.5*(xmin+xmax);
|
| 147 |
Double_t integ = lambda * TMath::Exp(-lambda*xmed) /
|
| 148 |
(TMath::Exp(-lambda*xmin)-TMath::Exp(-lambda*xmax));
|
| 149 |
|
| 150 |
return integ*TMath::Exp(-lambda*(mass-xmed));
|
| 151 |
}
|
| 152 |
|
| 153 |
//--------------------------------------------------------------------------------------------------
|
| 154 |
Double_t TPFuncs::ExpRange(Double_t *x, Double_t *par)
|
| 155 |
{
|
| 156 |
// Probability of an exponential in a given interval.
|
| 157 |
|
| 158 |
Double_t xx = x[0];
|
| 159 |
Double_t norm = par[0];
|
| 160 |
Double_t lambda = par[1];
|
| 161 |
Double_t xmin = par[2];
|
| 162 |
Double_t xmax = par[3];
|
| 163 |
|
| 164 |
Double_t ret = norm * ExpRange(xx,lambda,xmin,xmax);
|
| 165 |
return ret;
|
| 166 |
}
|
| 167 |
|
| 168 |
//--------------------------------------------------------------------------------------------------
|
| 169 |
Double_t TPFuncs::ZLineShapePlusBkg(Double_t *x, Double_t *par)
|
| 170 |
{
|
| 171 |
// Z line shape (BreitGaus) plus background (ExpRange).
|
| 172 |
|
| 173 |
// Parameters:
|
| 174 |
// 0: overall norm
|
| 175 |
// 1: relative fraction of signal and background
|
| 176 |
// 2: relative fraction of exponential and polynomial
|
| 177 |
// 3: most probable location) Breit mean
|
| 178 |
// 4: width (scale) Breit-Wigner
|
| 179 |
// 5: width (sigma) of convoluted Gaussian function
|
| 180 |
// 6: interference fraction between Z and gamma
|
| 181 |
// 7: lambda of exponential
|
| 182 |
// 8: minimum x (for normalization)
|
| 183 |
// 9: maximum x (for normalization)
|
| 184 |
|
| 185 |
Double_t xmin=par[8];
|
| 186 |
Double_t xmax=par[9];
|
| 187 |
Double_t s = (1.0 - par[1]) * BreitGaus(x[0], par[3], par[4], par[5], par[6], xmin, xmax);
|
| 188 |
Double_t b = par[1] * ((1.0 - par[2]) * ExpRange(x[0], par[7], xmin, xmax) + par[2]/(xmax-xmin));
|
| 189 |
|
| 190 |
return (s+b)*par[0];
|
| 191 |
}
|
| 192 |
|
| 193 |
//--------------------------------------------------------------------------------------------------
|
| 194 |
TF1 *TPFuncs::CreateZLineShapePlusBkg(Double_t norm, Double_t xmin, Double_t xmax, const char *n)
|
| 195 |
{
|
| 196 |
// Create TF1 for Z line shape and background.
|
| 197 |
|
| 198 |
TF1 *f = new TF1(n,mithep::TPFuncs::ZLineShapePlusBkg,xmin,xmax,10);
|
| 199 |
f->SetParName(0,"norm");
|
| 200 |
f->SetParName(1,"f_sb");
|
| 201 |
f->SetParName(2,"f_eb");
|
| 202 |
f->SetParName(3,"mass");
|
| 203 |
f->SetParName(4,"width");
|
| 204 |
f->SetParName(5,"gsig");
|
| 205 |
f->SetParName(6,"f_int");
|
| 206 |
f->SetParName(7,"lambda");
|
| 207 |
f->SetParName(8,"min");
|
| 208 |
f->SetParName(9,"max");
|
| 209 |
|
| 210 |
f->SetParameters(norm,0.5,0.5,91.,1.,1.,0.,0.1,xmin,xmax);
|
| 211 |
f->SetParLimits(0,0,1e6);
|
| 212 |
f->SetParLimits(1,0,1);
|
| 213 |
f->SetParLimits(2,0,1);
|
| 214 |
f->SetParLimits(3,xmin,xmax);
|
| 215 |
f->SetParLimits(4,0.1,3);
|
| 216 |
f->SetParLimits(5,0.1,3);
|
| 217 |
f->SetParLimits(6,0,0);
|
| 218 |
f->SetParLimits(7,0,10);
|
| 219 |
f->SetParLimits(8,xmin,xmin);
|
| 220 |
f->SetParLimits(9,xmax,xmax);
|
| 221 |
|
| 222 |
return f;
|
| 223 |
}
|
