Logo ROOT  

Reference Guide
 
Loading...
Searching...
No Matches
FitAwmi.C File Reference
Tutorials » Legacy tutorials » Spectrum tutorials

Detailed Description

This macro fits the source spectrum using the AWMI algorithm from the "TSpectrumFit" class ("TSpectrum" class is used to find peaks).

To try this macro, in a ROOT (5 or 6) prompt, do:

root > .x FitAwmi.C

or:

root > .x FitAwmi.C++
root > FitAwmi(); // re-run with another random set of peaks
created -9.7 39.8942 10
created -9.1 35.9048 9
created -8.5 15.9577 4
created -7.9 3.98942 1
created -7.3 15.9577 4
created -6.7 35.9048 9
created -6.1 31.9154 8
created -5.5 31.9154 8
created -4.9 31.9154 8
created -4.3 27.926 7
created -3.7 7.97885 2
created -3.1 7.97885 2
created -2.5 3.98942 1
created -1.9 31.9154 8
created -1.3 19.9471 5
created -0.7 31.9154 8
created -0.1 19.9471 5
created 0.5 23.9365 6
created 1.1 19.9471 5
created 1.7 15.9577 4
created 2.3 11.9683 3
created 2.9 23.9365 6
created 3.5 7.97885 2
created 4.1 35.9048 9
created 4.7 3.98942 1
created 5.3 7.97885 2
created 5.9 39.8942 10
created 6.5 3.98942 1
created 7.1 7.97885 2
created 7.7 3.98942 1
created 8.3 15.9577 4
created 8.9 7.97885 2
created 9.5 15.9577 4
the total number of created peaks = 33 with sigma = 0.1
the total number of found peaks = 33 with sigma = 0.100002 (+-3.34116e-05)
fit chi^2 = 2.8278e-06
found -9.7 (+-0.000233565) 39.8939 (+-0.0917909) 10.0001 (+-0.000753281)
found 5.9 (+-0.000231617) 39.8936 (+-0.0917091) 10 (+-0.00075261)
found -9.1 (+-0.000246286) 35.9048 (+-0.087092) 9.00018 (+-0.000714719)
found -6.7 (+-0.000246065) 35.9047 (+-0.0870819) 9.00015 (+-0.000714637)
found 4.1 (+-0.000244247) 35.9043 (+-0.0870068) 9.00003 (+-0.00071402)
found -6.1 (+-0.000261988) 31.9157 (+-0.0821406) 8.00022 (+-0.000674086)
found -5.5 (+-0.00026186) 31.9156 (+-0.0821355) 8.00021 (+-0.000674044)
found -4.9 (+-0.000261721) 31.9156 (+-0.0821299) 8.00019 (+-0.000673998)
found -1.9 (+-0.000259871) 31.9151 (+-0.0820607) 8.00008 (+-0.00067343)
found -0.7 (+-0.000260942) 31.9153 (+-0.0820993) 8.00013 (+-0.000673747)
found -4.3 (+-0.000278952) 27.9259 (+-0.0767982) 7.00013 (+-0.000630243)
found 0.5 (+-0.000301937) 23.9366 (+-0.0711184) 6.00013 (+-0.000583632)
found 2.9 (+-0.000300514) 23.9363 (+-0.0710775) 6.00006 (+-0.000583297)
found -1.3 (+-0.000332612) 19.9476 (+-0.0649692) 5.00021 (+-0.000533169)
found -0.100001 (+-0.000332174) 19.9475 (+-0.0649579) 5.00018 (+-0.000533077)
found 1.1 (+-0.000331194) 19.9473 (+-0.064933) 5.00013 (+-0.000532872)
found -8.5 (+-0.000370087) 15.9579 (+-0.0580772) 4.00013 (+-0.00047661)
found 1.7 (+-0.000370263) 15.9578 (+-0.0580775) 4.0001 (+-0.000476612)
found -7.3 (+-0.000370087) 15.9579 (+-0.0580772) 4.00013 (+-0.00047661)
found 8.3 (+-0.000367783) 15.9575 (+-0.0580297) 4.00004 (+-0.00047622)
found 9.5 (+-0.000365813) 15.9577 (+-0.0580002) 4.00008 (+-0.000475978)
found 2.3 (+-0.000429512) 11.9686 (+-0.0503269) 3.00013 (+-0.000413007)
found -3.70001 (+-0.00052706) 7.97916 (+-0.0411037) 2.00012 (+-0.000337317)
found 3.5 (+-0.000530831) 7.97947 (+-0.041144) 2.00019 (+-0.000337648)
found 5.30001 (+-0.000526953) 7.97927 (+-0.0411053) 2.00014 (+-0.00033733)
found 8.9 (+-0.000527079) 7.97911 (+-0.0411026) 2.0001 (+-0.000337308)
found -3.1 (+-0.000522415) 7.97886 (+-0.041055) 2.00004 (+-0.000336918)
found 7.1 (+-0.00052112) 7.9788 (+-0.0410424) 2.00003 (+-0.000336814)
found 6.49998 (+-0.00075357) 3.98998 (+-0.0291118) 1.00016 (+-0.000238905)
found 4.69998 (+-0.000752903) 3.98992 (+-0.0291076) 1.00014 (+-0.000238871)
found -7.9 (+-0.000751472) 3.98977 (+-0.0290974) 1.0001 (+-0.000238787)
found -2.49999 (+-0.000752181) 3.98987 (+-0.0291031) 1.00013 (+-0.000238834)
found 7.7 (+-0.000748419) 3.98966 (+-0.0290806) 1.00008 (+-0.00023865)
#include "TROOT.h"
#include "TMath.h"
#include "TRandom.h"
#include "TH1.h"
#include "TF1.h"
#include "TCanvas.h"
#include "TSpectrum.h"
#include "TSpectrumFit.h"
#include "TPolyMarker.h"
#include "TList.h"
#include <iostream>
TH1F *FitAwmi_Create_Spectrum(void)
{
Int_t nbins = 1000;
Double_t xmin = -10., xmax = 10.;
delete gROOT->FindObject("h"); // prevent "memory leak"
TH1F *h = new TH1F("h", "simulated spectrum", nbins, xmin, xmax);
h->SetStats(kFALSE);
TF1 f("f", "TMath::Gaus(x, [0], [1], 1)", xmin, xmax);
// f.SetParNames("mean", "sigma");
gRandom->SetSeed(0); // make it really random
// create well separated peaks with exactly known means and areas
// note: TSpectrumFit assumes that all peaks have the same sigma
Double_t sigma = (xmax - xmin) / Double_t(nbins) * Int_t(gRandom->Uniform(2., 6.));
Int_t npeaks = 0;
while (xmax > (xmin + 6. * sigma)) {
npeaks++;
xmin += 3. * sigma; // "mean"
f.SetParameters(xmin, sigma);
Double_t area = 1. * Int_t(gRandom->Uniform(1., 11.));
h->Add(&f, area, ""); // "" ... or ... "I"
std::cout << "created " << xmin << " " << (area / sigma / TMath::Sqrt(TMath::TwoPi())) << " " << area
<< std::endl;
xmin += 3. * sigma;
}
std::cout << "the total number of created peaks = " << npeaks << " with sigma = " << sigma << std::endl;
return h;
}
void FitAwmi(void)
{
TH1F *h = FitAwmi_Create_Spectrum();
TCanvas *cFit = ((TCanvas *)(gROOT->GetListOfCanvases()->FindObject("cFit")));
if (!cFit)
cFit = new TCanvas("cFit", "cFit", 10, 10, 1000, 700);
else
cFit->Clear();
h->Draw("L");
Int_t i, nfound, bin;
Int_t nbins = h->GetNbinsX();
Double_t *source = new Double_t[nbins];
Double_t *dest = new Double_t[nbins];
for (i = 0; i < nbins; i++)
source[i] = h->GetBinContent(i + 1);
TSpectrum *s = new TSpectrum(); // note: default maxpositions = 100
// searching for candidate peaks positions
nfound = s->SearchHighRes(source, dest, nbins, 2., 2., kFALSE, 10000, kFALSE, 0);
// filling in the initial estimates of the input parameters
Bool_t *FixPos = new Bool_t[nfound];
Bool_t *FixAmp = new Bool_t[nfound];
for (i = 0; i < nfound; i++)
FixAmp[i] = FixPos[i] = kFALSE;
Double_t *Pos, *Amp = new Double_t[nfound]; // ROOT 6
Pos = s->GetPositionX(); // 0 ... (nbins - 1)
for (i = 0; i < nfound; i++) {
bin = 1 + Int_t(Pos[i] + 0.5); // the "nearest" bin
Amp[i] = h->GetBinContent(bin);
}
TSpectrumFit *pfit = new TSpectrumFit(nfound);
pfit->SetFitParameters(0, (nbins - 1), 1000, 0.1, pfit->kFitOptimChiCounts, pfit->kFitAlphaHalving, pfit->kFitPower2,
pfit->kFitTaylorOrderFirst);
pfit->SetPeakParameters(2., kFALSE, Pos, FixPos, Amp, FixAmp);
// pfit->SetBackgroundParameters(source[0], kFALSE, 0., kFALSE, 0., kFALSE);
pfit->FitAwmi(source);
Double_t *Positions = pfit->GetPositions();
Double_t *PositionsErrors = pfit->GetPositionsErrors();
Double_t *Amplitudes = pfit->GetAmplitudes();
Double_t *AmplitudesErrors = pfit->GetAmplitudesErrors();
Double_t *Areas = pfit->GetAreas();
Double_t *AreasErrors = pfit->GetAreasErrors();
delete gROOT->FindObject("d"); // prevent "memory leak"
TH1F *d = new TH1F(*h);
d->SetNameTitle("d", "");
d->Reset("M");
for (i = 0; i < nbins; i++)
d->SetBinContent(i + 1, source[i]);
Double_t x1 = d->GetBinCenter(1), dx = d->GetBinWidth(1);
Double_t sigma, sigmaErr;
pfit->GetSigma(sigma, sigmaErr);
// current TSpectrumFit needs a sqrt(2) correction factor for sigma
sigma /= TMath::Sqrt2();
sigmaErr /= TMath::Sqrt2();
// convert "bin numbers" into "x-axis values"
sigma *= dx;
sigmaErr *= dx;
std::cout << "the total number of found peaks = " << nfound << " with sigma = " << sigma << " (+-" << sigmaErr << ")"
<< std::endl;
std::cout << "fit chi^2 = " << pfit->GetChi() << std::endl;
for (i = 0; i < nfound; i++) {
bin = 1 + Int_t(Positions[i] + 0.5); // the "nearest" bin
Pos[i] = d->GetBinCenter(bin);
Amp[i] = d->GetBinContent(bin);
// convert "bin numbers" into "x-axis values"
Positions[i] = x1 + Positions[i] * dx;
PositionsErrors[i] *= dx;
Areas[i] *= dx;
AreasErrors[i] *= dx;
std::cout << "found " << Positions[i] << " (+-" << PositionsErrors[i] << ") " << Amplitudes[i] << " (+-"
<< AmplitudesErrors[i] << ") " << Areas[i] << " (+-" << AreasErrors[i] << ")" << std::endl;
}
d->SetLineColor(kRed);
d->SetLineWidth(1);
d->Draw("SAME L");
TPolyMarker *pm = ((TPolyMarker *)(h->GetListOfFunctions()->FindObject("TPolyMarker")));
if (pm) {
h->GetListOfFunctions()->Remove(pm);
delete pm;
}
pm = new TPolyMarker(nfound, Pos, Amp);
h->GetListOfFunctions()->Add(pm);
pm->SetMarkerStyle(23);
pm->SetMarkerColor(kRed);
pm->SetMarkerSize(1);
// cleanup
delete pfit;
delete[] Amp;
delete[] FixAmp;
delete[] FixPos;
delete s;
delete[] dest;
delete[] source;
return;
}
d
#define d(i)
Definition RSha256.hxx:102
f
#define f(i)
Definition RSha256.hxx:104
h
#define h(i)
Definition RSha256.hxx:106
Bool_t
bool Bool_t
Definition RtypesCore.h:63
Int_t
int Int_t
Definition RtypesCore.h:45
kFALSE
constexpr Bool_t kFALSE
Definition RtypesCore.h:94
Double_t
double Double_t
Definition RtypesCore.h:59
kRed
@ kRed
Definition Rtypes.h:66
TRangeDynCast
ROOT::Detail::TRangeCast< T, true > TRangeDynCast
TRangeDynCast is an adapter class that allows the typed iteration through a TCollection.
Definition TCollection.h:358
dest
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t dest
Definition TGWin32VirtualXProxy.cxx:164
x1
Option_t Option_t TPoint TPoint const char x1
Definition TGWin32VirtualXProxy.cxx:70
xmin
float xmin
Definition THbookFile.cxx:95
xmax
float xmax
Definition THbookFile.cxx:95
gROOT
#define gROOT
Definition TROOT.h:414
gRandom
R__EXTERN TRandom * gRandom
Definition TRandom.h:62
TCanvas
The Canvas class.
Definition TCanvas.h:23
TF1
1-Dim function class
Definition TF1.h:234
TH1F
1-D histogram with a float per channel (see TH1 documentation)
Definition TH1.h:877
TPolyMarker
A PolyMarker is defined by an array on N points in a 2-D space.
Definition TPolyMarker.h:31
TRandom::SetSeed
virtual void SetSeed(ULong_t seed=0)
Set the random generator seed.
Definition TRandom.cxx:615
TRandom::Uniform
virtual Double_t Uniform(Double_t x1=1)
Returns a uniform deviate on the interval (0, x1).
Definition TRandom.cxx:682
TSpectrumFit
Advanced 1-dimensional spectra fitting functions.
Definition TSpectrumFit.h:18
TSpectrum
Advanced Spectra Processing.
Definition TSpectrum.h:18
TSpectrum::SearchHighRes
Int_t SearchHighRes(Double_t *source, Double_t *destVector, Int_t ssize, Double_t sigma, Double_t threshold, bool backgroundRemove, Int_t deconIterations, bool markov, Int_t averWindow)
One-dimensional high-resolution peak search function.
Definition TSpectrum.cxx:2119
TSpectrum::GetPositionX
Double_t * GetPositionX() const
Definition TSpectrum.h:58
sigma
const Double_t sigma
Definition h1analysisProxy.h:11
TMath::Sqrt2
constexpr Double_t Sqrt2()
Definition TMath.h:86
TMath::Sqrt
Double_t Sqrt(Double_t x)
Returns the square root of x.
Definition TMath.h:666
TMath::TwoPi
constexpr Double_t TwoPi()
Definition TMath.h:44
Author

Definition in file FitAwmi.C.