| 1 |
//******************************************//
|
| 2 |
// author:
|
| 3 |
//
|
| 4 |
// This function evaluates the effective
|
| 5 |
// sigma of a histogram.
|
| 6 |
//
|
| 7 |
//
|
| 8 |
// TODO list:
|
| 9 |
//
|
| 10 |
//
|
| 11 |
//
|
| 12 |
// Modified by:
|
| 13 |
//
|
| 14 |
//******************************************//
|
| 15 |
|
| 16 |
|
| 17 |
|
| 18 |
|
| 19 |
//Double_t effSigma(TH1 * hist)
|
| 20 |
|
| 21 |
void effSigma(TH1 * hist,double res[])
|
| 22 |
|
| 23 |
{
|
| 24 |
|
| 25 |
TAxis *xaxis = hist->GetXaxis();
|
| 26 |
Int_t nb = xaxis->GetNbins();
|
| 27 |
if(nb < 10) {
|
| 28 |
cout << "effsigma: Not a valid histo. nbins = " << nb << endl;
|
| 29 |
return ;
|
| 30 |
}
|
| 31 |
|
| 32 |
Double_t bwid = xaxis->GetBinWidth(1);
|
| 33 |
if(bwid == 0) {
|
| 34 |
cout << "effsigma: Not a valid histo. bwid = " << bwid << endl;
|
| 35 |
return ;
|
| 36 |
}
|
| 37 |
Double_t xmax = xaxis->GetXmax();
|
| 38 |
Double_t xmin = xaxis->GetXmin();
|
| 39 |
Double_t ave = hist->GetMean();
|
| 40 |
Double_t rms = hist->GetRMS();
|
| 41 |
|
| 42 |
Double_t total=0.;
|
| 43 |
for(Int_t i=0; i<nb+2; i++) {
|
| 44 |
total+=hist->GetBinContent(i);
|
| 45 |
}
|
| 46 |
// if(total < 100.) {
|
| 47 |
if(total < 50.) {
|
| 48 |
cout << "effsigma: Too few entries " << total << endl;
|
| 49 |
return ;
|
| 50 |
}
|
| 51 |
Int_t ierr=0;
|
| 52 |
Int_t ismin=999;
|
| 53 |
|
| 54 |
Double_t rlim=0.683*total;
|
| 55 |
Int_t nrms=int(rms/(bwid)); // Set scan size to +/- rms
|
| 56 |
if(nrms > nb/10) nrms=nb/10; // Could be tuned...
|
| 57 |
|
| 58 |
Double_t widmin=9999999.;
|
| 59 |
|
| 60 |
// cout<<"iscan: "<< xmin <<" "<< xmax <<" "<< ave <<" "<< rms <<" "<< rlim <<" "<< nrms <<endl;
|
| 61 |
|
| 62 |
int binlow = 1;
|
| 63 |
int binhigh = nb;
|
| 64 |
|
| 65 |
double mean_min = ave;
|
| 66 |
|
| 67 |
int binlow2 = 1;
|
| 68 |
int binhigh2 = nb;
|
| 69 |
|
| 70 |
|
| 71 |
|
| 72 |
for(Int_t iscan=-nrms;iscan<nrms+1;iscan++) { // Scan window centre
|
| 73 |
Int_t ibm=(ave-xmin)/bwid+1+iscan;
|
| 74 |
Double_t x=(ibm-0.5)*bwid+xmin;
|
| 75 |
Double_t xj=x;
|
| 76 |
Double_t xk=x;
|
| 77 |
Int_t jbm=ibm;
|
| 78 |
Int_t kbm=ibm;
|
| 79 |
Double_t bin=hist->GetBinContent(ibm);
|
| 80 |
total=bin;
|
| 81 |
|
| 82 |
// cout<<"ibm: "<< iscan <<" "<< ibm <<" "<< xj<<" "<< xk <<" "<<jbm <<" "<<kbm <<endl;
|
| 83 |
|
| 84 |
double mean = bin * x;
|
| 85 |
|
| 86 |
|
| 87 |
for(Int_t j=1;j<nb;j++){
|
| 88 |
if(jbm < nb) {
|
| 89 |
jbm++;
|
| 90 |
xj+=bwid;
|
| 91 |
bin=hist->GetBinContent(jbm);
|
| 92 |
total+=bin;
|
| 93 |
|
| 94 |
if(total > rlim) break;
|
| 95 |
}
|
| 96 |
else ierr=1;
|
| 97 |
if(kbm > 0) {
|
| 98 |
kbm--;
|
| 99 |
xk-=bwid;
|
| 100 |
bin=hist->GetBinContent(kbm);
|
| 101 |
total+=bin;
|
| 102 |
|
| 103 |
if(total > rlim) break;
|
| 104 |
}
|
| 105 |
else ierr=1;
|
| 106 |
}
|
| 107 |
Double_t dxf=(total-rlim)*bwid/bin;
|
| 108 |
Double_t wid=(xj-xk+bwid-dxf)*0.5;
|
| 109 |
|
| 110 |
|
| 111 |
//cout<<"dxf: "<< total <<" "<< rlim <<" "<< bwid <<" "<< bin <<" "<< wid <<" "<< xj<<" "<<xk<<" "<< dxf <<endl;
|
| 112 |
|
| 113 |
if(wid < widmin) {
|
| 114 |
widmin=wid;
|
| 115 |
ismin=iscan;
|
| 116 |
|
| 117 |
binlow = kbm;
|
| 118 |
binhigh = jbm;
|
| 119 |
|
| 120 |
|
| 121 |
binlow2 = ibm - 2*(ibm - kbm);
|
| 122 |
binhigh2 = ibm + 2*(-ibm + jbm);
|
| 123 |
|
| 124 |
}
|
| 125 |
|
| 126 |
// cout<<"widmin: "<< wid<<" "<< widmin <<" "<< ismin <<" "<<ibm <<" "<< jbm<<" "<< kbm <<" "<<x <<" "<< xj <<" "<< xk <<" "<<(ibm-0.5)*bwid+xmin<<" "<<
|
| 127 |
//" "<< (jbm-0.5)*bwid+xmin <<" "<<(kbm-0.5)*bwid+xmin <<endl;
|
| 128 |
|
| 129 |
}
|
| 130 |
if(ismin == nrms || ismin == -nrms) ierr=3;
|
| 131 |
if(ierr != 0) cout << "effsigma: Error of type " << ierr << endl;
|
| 132 |
|
| 133 |
//return widmin;
|
| 134 |
|
| 135 |
double mean = 0;
|
| 136 |
double n =0;
|
| 137 |
for(int b = binlow ; b<= binhigh; b++){
|
| 138 |
|
| 139 |
if( b<1 || b> nb) continue;
|
| 140 |
|
| 141 |
double x= (b-0.5)*bwid+xmin;
|
| 142 |
mean += hist->GetBinContent(b) * x;
|
| 143 |
n += hist->GetBinContent(b);
|
| 144 |
}
|
| 145 |
|
| 146 |
mean /= n;
|
| 147 |
|
| 148 |
double mean2 = 0;
|
| 149 |
double n2 =0;
|
| 150 |
for(int b = binlow2 ; b<= binhigh2; b++){
|
| 151 |
|
| 152 |
if( b<1 || b> nb) continue;
|
| 153 |
|
| 154 |
double x= (b-0.5)*bwid+xmin;
|
| 155 |
mean2 += hist->GetBinContent(b) * x;
|
| 156 |
n2 += hist->GetBinContent(b);
|
| 157 |
}
|
| 158 |
|
| 159 |
mean2 /= n2;
|
| 160 |
|
| 161 |
|
| 162 |
//cout<<widmin <<" "<<mean <<endl;
|
| 163 |
|
| 164 |
res[0] = mean;
|
| 165 |
res[1] = mean2;
|
| 166 |
|
| 167 |
res[2] = widmin;
|
| 168 |
|
| 169 |
|
| 170 |
}
|
