• Created by Unknown User (ddavis), last modified on Apr 18, 2005

gtlikelihood
************
Overall the likelihood tool is not very intuitive. For instance it would be nice to see the
starting statistic and see that the likelihood statistic is changing with each interation.

Terms like FCN= 15883.48 and EDM= 0.17E+01 should be discussed in the documentation (I
couldn't find them).

It would be useful if likelihood could output the current 2-D model or a residual map.

It would be nice if the parameters were listed in the same order (where possible)
as for the previous tools.

>plist gtlikelihood
Parameters for /home/lhea3/ddavis/pfiles/gtlikelihood.par
rspfunc = TEST Response functions to use (enumlist=DC1)
(use_energy_dispersion = no) Use energy dispersion?
source_model_file = virgo_model.xml Source model file
source_model_output_file = virgo_model_out.xml Source model output file
flux_style_model_file = virgo_flux_model.xml flux-style output file name
statistic = BINNED Statistic to use (enumlist=[BINNED|UNBINNED|OPTEM])
optimizer = MINUIT Optimizer (enumlist=[LBFGS|MINUIT|DRMNGB])
fit_tolerance = 0.0001 Fit tolerance
(find_Ts_mins = no) Re-optimize for TS fits?
query_for_refit = yes Allow for refitting?
evfile = ../L050405162511FE57EF3F19_PH00.fits Event file
scfile = ../FT2_checkout2_scData_0000.fits Spacecraft file
exposure_map_file = Virgo_expmap.fits Exposure file
(plot = no) Plot unbinned counts spectra?
counts_map_file = ./n4486_region_img.fits Counts map file
exposure_cube_file = ./virgo_livetimecube.fits Exposure hypercube file
binned_exposure_map = ./Virgo_expmap.fits Binned exposure map
(chatter = 2) Output verbosity (min=0, max=4)
(clobber = yes) Overwrite existing output files
(mode = ql) Mode of automatic parameters

Binned Likelihood fails with

time ST/bin/gtlikelihood mode=h
Caught N3tip12TipExceptionE at the top level: Could not open FITS extension "./Virgo_expmap.fits[Energies]" (CFITSIO ERROR 301: illegal HDU number)
1.230u 0.110s 0:01.65 81.2% 0+0k 0+0io 4075pf+0w

try UNBINNED

plist gtlikelihood
Parameters for /home/lhea3/ddavis/pfiles/gtlikelihood.par
rspfunc = TEST Response functions to use (enumlist=DC1)
(use_energy_dispersion = no) Use energy dispersion?
source_model_file = virgo_model.xml Source model file
source_model_output_file = virgo_model_out.xml Source model output file
flux_style_model_file = virgo_flux_model.xml flux-style output file name
statistic = UNBINNED Statistic to use (enumlist=BINNED)
optimizer = MINUIT Optimizer (enumlist=LBFGS)
fit_tolerance = 0.0001 Fit tolerance
(find_Ts_mins = no) Re-optimize for TS fits?
query_for_refit = yes Allow for refitting?
evfile = ../L050405162511FE57EF3F19_PH00.fits Event file
scfile = ../FT2_checkout2_scData_0000.fits Spacecraft file
exposure_map_file = Virgo_expmap.fits Exposure file
(plot = no) Plot unbinned counts spectra?
counts_map_file = ./n4486_region_img.fits Counts map file
exposure_cube_file = ./virgo_livetimecube.fits Exposure hypercube file
binned_exposure_map = ./Virgo_expmap.fits Binned exposure map
(chatter = 2) Output verbosity (min=0, max=4)
(clobber = yes) Overwrite existing output files
(mode = ql) Mode of automatic parameters

>time ST/bin/gtlikelihood mode=h
n
**********

    • 1 **SET PRINT 0.000
      **********
      **********
    • 2 **SET NOWARN
      **********

PARAMETER DEFINITIONS:
NO. NAME VALUE STEP SIZE LIMITS
1 'Prefactor ' 1.6000 1.0000 0.10000E-04 100.00
2 'Prefactor ' 11.000 1.0000 0.10000E-02 1000.0
3 'Prefactor ' 10.000 1.0000 0.10000E-02 1000.0
4 'Index ' -2.1000 1.0000 -5.0000 -1.0000
5 'Prefactor ' 10.000 1.0000 0.10000E-02 1000.0
6 'Index ' -2.0000 1.0000 -5.0000 -1.0000
**********

    • 3 **SET ERR 0.5000
      **********
      **********
    • 4 **SET GRAD 1.000
      **********
      **********
    • 5 **MIGRAD 200.0 0.5384E+05
      **********

MIGRAD MINIMIZATION HAS CONVERGED.

MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.

FCN= 266007.6 FROM MIGRAD STATUS=CONVERGED 108 CALLS 109 TOTAL
EDM= 0.23E+01 STRATEGY= 1 ERR MATRIX NOT POS-DEF

EXT PARAMETER APPROXIMATE STEP FIRST
NO. NAME VALUE ERROR SIZE DERIVATIVE
1 Prefactor 2.7712 0.16150E-01 0.14687E-01 133.85
2 Prefactor 8.2710 0.31389 0.95978E-02 76.269
3 Prefactor 0.23982 0.39998 0.10063E-01 17.238
4 Index -1.6842 0.50745 0.33823 -3.5363
5 Prefactor 4.9086 0.81515E-01 0.10063E-01 -56.324
6 Index -1.7814 0.74325E-02 0.78442E-01 -0.23281
ERR DEF= 0.500
**********

    • 6 **HESSE
      **********

FCN= 266007.6 FROM HESSE STATUS=OK 44 CALLS 153 TOTAL
EDM= 0.14E+01 STRATEGY= 1 ERROR MATRIX ACCURATE

EXT PARAMETER INTERNAL INTERNAL
NO. NAME VALUE ERROR STEP SIZE VALUE
1 Prefactor 2.7712 0.13809 0.64907E-04 -1.2363
2 Prefactor 8.2710 2.6509 0.22599E-03 -1.3887
3 Prefactor 0.23982 0.48673E-01 0.10772E-03 -1.5399
4 Index -1.6842 0.65888E-01 0.14321E-01 0.71803
5 Prefactor 4.9086 0.17193 0.41991E-03 -1.4306
6 Index -1.7814 0.15526E-01 0.33972E-03 0.65515
ERR DEF= 0.500
Final values:
Prefactor = 2.77117
Prefactor = 8.27103
Prefactor = 0.239821
Index = -1.68419
Prefactor = 4.90856
Index = -1.78144
Minuit fit quality: 3 estimated distance: 1.36053
Minuit parameter uncertainties:
1 0.138093
2 2.65127
3 0.0486731
4 0.0659086
5 0.171929
6 0.0155259
Computing TS values for each source (4 total)
....!

Extragalactic Diffuse:
Prefactor: 2.77117 +/- 0.138093
Index: -2.1
Scale: 100
Npred: 21008.5

Galactic Diffuse:
Prefactor: 8.27103 +/- 2.65127
Index: -2.1
Scale: 100
Npred: 3236.53

src 1:
Prefactor: 0.239821 +/- 0.0486731
Index: 1.68419 +/ 0.0659086
Scale: 100
Npred: 178.194
ROI distance: 9.08485
TS value: 114.794

src 2:
Prefactor: 4.90856 +/- 0.171929
Index: 1.78144 +/ 0.0155259
Scale: 100
Npred: 3030.64
ROI distance: 10.2176
TS value: 6456.96

-log(Likelihood): 266008

Writing fitted model to virgo_model_out.xml
Refit? y
Writing flux-style xml model file to virgo_flux_model.xml
2546.680u 10.750s 42:56.97 99.2% 0+0k 0+0io 4294pf+0w

42min for one iteration

*******
gttsmap
*******

time ST/bin/gttsmap mode=h
....................!
76344.490u 120.520s 21:34:26.99 -12.-1% 0+0k 0+0io 4285pf+0w

  • No labels