# rf403_weightedevts¶

'DATA AND CATEGORIES' RooFit tutorial macro #403

Using weights in unbinned datasets

Author: Clemens Lange, Wouter Verkerke (C version)
This notebook tutorial was automatically generated with ROOTBOOK-izer from the macro found in the ROOT repository on Wednesday, November 30, 2022 at 11:23 AM.

In [1]:
from __future__ import print_function
import ROOT

Welcome to JupyROOT 6.27/01


## Create observable and unweighted dataset¶

Declare observable

In [2]:
x = ROOT.RooRealVar("x", "x", -10, 10)
x.setBins(40)


Construction a uniform pdf

In [3]:
p0 = ROOT.RooPolynomial("px", "px", x)


Sample 1000 events from pdf

In [4]:
data = p0.generate({x}, 1000)


## Calculate weight and make dataset weighted¶

Construct formula to calculate (fake) weight for events

In [5]:
wFunc = ROOT.RooFormulaVar("w", "event weight", "(x*x+10)", [x])


Add column with variable w to previously generated dataset

In [6]:
w = data.addColumn(wFunc)


Dataset d is now a dataset with two observable (x,w) with 1000 entries

In [7]:
data.Print()

RooDataSet::pxData[x,w] = 1000 entries


Instruct dataset wdata in interpret w as event weight rather than as observable

In [8]:
wdata = ROOT.RooDataSet(data.GetName(), data.GetTitle(), data, data.get(), "", w.GetName())


Dataset d is now a dataset with one observable (x) with 1000 entries and a sum of weights of ~430K

In [9]:
wdata.Print()

RooDataSet::pxData[x,weight:w] = 1000 entries (43238.9 weighted)


## Unbinned ML fit to weighted data¶

Construction quadratic polynomial pdf for fitting

In [10]:
a0 = ROOT.RooRealVar("a0", "a0", 1)
a1 = ROOT.RooRealVar("a1", "a1", 0, -1, 1)
a2 = ROOT.RooRealVar("a2", "a2", 1, 0, 10)
p2 = ROOT.RooPolynomial("p2", "p2", x, [a0, a1, a2], 0)


Fit quadratic polynomial to weighted data

NOTE: A plain Maximum likelihood fit to weighted data does in general NOT result in correct error estimates, individual event weights represent Poisson statistics themselves.

Fit with 'wrong' errors

In [11]:
r_ml_wgt = p2.fitTo(wdata, Save=True)

[#0] WARNING:InputArguments -- RooAbsPdf::fitTo(p2) WARNING: a likelihood fit is requested of what appears to be weighted data.
While the estimated values of the parameters will always be calculated taking the weights into account,
there are multiple ways to estimate the errors of the parameters. You are advised to make an
explicit choice for the error calculation:
- Either provide SumW2Error(true), to calculate a sum-of-weights-corrected HESSE error matrix
(error will be proportional to the number of events in MC).
- Or provide SumW2Error(false), to return errors from original HESSE error matrix
(which will be proportional to the sum of the weights, i.e., a dataset with <sum of weights> events).
- Or provide AsymptoticError(true), to use the asymptotically correct expression
(for details see https://arxiv.org/abs/1911.01303).
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: activating const optimization
**********
**    1 **SET PRINT           1
**********
**********
**********
PARAMETER DEFINITIONS:
NO.   NAME         VALUE      STEP SIZE      LIMITS
1 a1           0.00000e+00  2.00000e-01   -1.00000e+00  1.00000e+00
2 a2           1.00000e+00  5.00000e-01    0.00000e+00  1.00000e+01
**********
**    3 **SET ERR         0.5
**********
**********
**    4 **SET PRINT           1
**********
**********
**    5 **SET STR           1
**********
NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
**********
**********
FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
FCN=119822 FROM MIGRAD    STATUS=INITIATE       25 CALLS          26 TOTAL
EDM= unknown      STRATEGY= 1      NO ERROR MATRIX
EXT PARAMETER               CURRENT GUESS       STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1           0.00000e+00   2.00000e-01   0.00000e+00   1.68737e+02
2  a2           1.52215e-01   5.00000e-01  -3.96120e-01   4.87454e+03
ERR DEF= 0.5
MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=119682 FROM MIGRAD    STATUS=CONVERGED      63 CALLS          64 TOTAL
EDM=1.29222e-07    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                   STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -4.85603e-03   4.03455e-03   9.64266e-04  -8.28728e-02
2  a2           9.86514e-02   2.41306e-03   5.83664e-04  -5.42318e-02
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
1.628e-05 -1.173e-08
-1.173e-08  5.823e-06
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00120   1.000 -0.001
2  0.00120  -0.001  1.000
**********
**    7 **SET ERR         0.5
**********
**********
**    8 **SET PRINT           1
**********
**********
**    9 **HESSE        1000
**********
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=119682 FROM HESSE     STATUS=OK             10 CALLS          74 TOTAL
EDM=1.28662e-07    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                INTERNAL      INTERNAL
NO.   NAME      VALUE            ERROR       STEP SIZE       VALUE
1  a1          -4.85603e-03   4.03459e-03   1.92853e-04  -4.85605e-03
2  a2           9.86514e-02   2.41310e-03   2.33465e-05  -1.37182e+00
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
1.628e-05 -4.575e-08
-4.575e-08  5.823e-06
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00470   1.000 -0.005
2  0.00470  -0.005  1.000
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: deactivating const optimization


A first order correction to estimated parameter errors in an (unbinned) ML fit can be obtained by calculating the covariance matrix as

V' = V C-1 V

where V is the covariance matrix calculated from a fit to -logL = - sum [ w_i log f(x_i) ] and C is the covariance matrix calculated from -logL' = -sum [ w_i^2 log f(x_i) ] (i.e. the weights are applied squared)

A fit in self mode can be performed as follows:

In [12]:
r_ml_wgt_corr = p2.fitTo(wdata, Save=True, SumW2Error=True)

[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: activating const optimization
**********
**   10 **SET PRINT           1
**********
**********
**********
PARAMETER DEFINITIONS:
NO.   NAME         VALUE      STEP SIZE      LIMITS
1 a1          -4.85603e-03  4.03459e-03   -1.00000e+00  1.00000e+00
2 a2           9.86514e-02  2.41310e-03    0.00000e+00  1.00000e+01
**********
**   12 **SET ERR         0.5
**********
**********
**   13 **SET PRINT           1
**********
**********
**   14 **SET STR           1
**********
NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
**********
**********
FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
FCN=119682 FROM MIGRAD    STATUS=INITIATE        4 CALLS           5 TOTAL
EDM= unknown      STRATEGY= 1      NO ERROR MATRIX
EXT PARAMETER               CURRENT GUESS       STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -4.85603e-03   4.03459e-03   4.03465e-03  -8.32948e-02
2  a2           9.86514e-02   2.41310e-03   2.44179e-03  -2.65748e-01
ERR DEF= 0.5
MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=119682 FROM MIGRAD    STATUS=CONVERGED      23 CALLS          24 TOTAL
EDM=5.53437e-08    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                   STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -4.85553e-03   4.03456e-03   9.63830e-04  -5.18346e-02
2  a2           9.86520e-02   2.41308e-03   5.83268e-04   4.40227e-02
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
1.628e-05 -1.171e-08
-1.171e-08  5.823e-06
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00120   1.000 -0.001
2  0.00120  -0.001  1.000
**********
**   16 **SET ERR         0.5
**********
**********
**   17 **SET PRINT           1
**********
**********
**   18 **HESSE        1000
**********
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=119682 FROM HESSE     STATUS=OK             10 CALLS          34 TOTAL
EDM=5.56466e-08    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                INTERNAL      INTERNAL
NO.   NAME      VALUE            ERROR       STEP SIZE       VALUE
1  a1          -4.85553e-03   4.03461e-03   1.92766e-04  -4.85555e-03
2  a2           9.86520e-02   2.41312e-03   2.33307e-05  -1.37182e+00
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
1.628e-05 -4.571e-08
-4.571e-08  5.823e-06
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00470   1.000 -0.005
2  0.00470  -0.005  1.000
[#1] INFO:Fitting -- RooAbsPdf::fitTo(p2) Calculating sum-of-weights-squared correction matrix for covariance matrix
**********
**   19 **SET ERR         0.5
**********
**********
**   20 **SET PRINT           1
**********
**********
**   21 **HESSE        1000
**********
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=6.84247e+06 FROM HESSE     STATUS=OK             14 CALLS          48 TOTAL
EDM=158914    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                INTERNAL      INTERNAL
NO.   NAME      VALUE            ERROR       STEP SIZE       VALUE
1  a1          -4.85553e-03   5.43362e-04   3.93243e-03  -4.85555e-03
2  a2           9.86520e-02   1.95133e-04   4.75947e-04  -1.37182e+00
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
2.952e-07  6.228e-09
6.228e-09  3.808e-08
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.05874   1.000  0.059
2  0.05874   0.059  1.000
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: deactivating const optimization


## Plot weighted data and fit result¶

Construct plot frame

In [13]:
frame = x.frame(Title="Unbinned ML fit, chi^2 fit to weighted data")


Plot data using sum-of-weights-squared error rather than Poisson errors

In [14]:
wdata.plotOn(frame, DataError="SumW2")

Out[14]:
<cppyy.gbl.RooPlot object at 0x9133840>

Overlay result of 2nd order polynomial fit to weighted data

In [15]:
p2.plotOn(frame)

Out[15]:
<cppyy.gbl.RooPlot object at 0x9133840>

## ML fit of pdf to equivalent unweighted dataset¶

Construct a pdf with the same shape as p0 after weighting

In [16]:
genPdf = ROOT.RooGenericPdf("genPdf", "x*x+10", [x])


Sample a dataset with the same number of events as data

In [17]:
data2 = genPdf.generate({x}, 1000)

[#1] INFO:NumericIntegration -- RooRealIntegral::init(genPdf_Int[x]) using numeric integrator RooIntegrator1D to calculate Int(x)
[#1] INFO:NumericIntegration -- RooRealIntegral::init(genPdf_Int[x]) using numeric integrator RooIntegrator1D to calculate Int(x)


Sample a dataset with the same number of weights as data

In [18]:
data3 = genPdf.generate({x}, 43000)

[#1] INFO:NumericIntegration -- RooRealIntegral::init(genPdf_Int[x]) using numeric integrator RooIntegrator1D to calculate Int(x)
[#1] INFO:NumericIntegration -- RooRealIntegral::init(genPdf_Int[x]) using numeric integrator RooIntegrator1D to calculate Int(x)


Fit the 2nd order polynomial to both unweighted datasets and save the results for comparison

In [19]:
r_ml_unw10 = p2.fitTo(data2, Save=True)
r_ml_unw43 = p2.fitTo(data3, Save=True)

[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: activating const optimization
**********
**   22 **SET PRINT           1
**********
**********
**********
PARAMETER DEFINITIONS:
NO.   NAME         VALUE      STEP SIZE      LIMITS
1 a1          -4.85553e-03  3.00247e-02   -1.00000e+00  1.00000e+00
2 a2           9.86520e-02  2.98987e-02    0.00000e+00  1.00000e+01
**********
**   24 **SET ERR         0.5
**********
**********
**   25 **SET PRINT           1
**********
**********
**   26 **SET STR           1
**********
NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
**********
**********
FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
FCN=2766.64 FROM MIGRAD    STATUS=INITIATE        6 CALLS           7 TOTAL
EDM= unknown      STRATEGY= 1      NO ERROR MATRIX
EXT PARAMETER               CURRENT GUESS       STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -4.85553e-03   3.00247e-02   3.00296e-02  -1.97827e+01
2  a2           9.86520e-02   2.98987e-02   3.06116e-02  -1.19823e+01
ERR DEF= 0.5
MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=2766.49 FROM MIGRAD    STATUS=CONVERGED      29 CALLS          30 TOTAL
EDM=7.62783e-07    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                   STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1           8.94396e-03   2.69244e-02   9.81558e-04   3.18558e-02
2  a2           1.01291e-01   1.67455e-02   6.10755e-04  -9.73691e-03
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
7.251e-04 -6.220e-07
-6.220e-07  2.804e-04
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00138   1.000 -0.001
2  0.00138  -0.001  1.000
**********
**   28 **SET ERR         0.5
**********
**********
**   29 **SET PRINT           1
**********
**********
**   30 **HESSE        1000
**********
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=2766.49 FROM HESSE     STATUS=OK             10 CALLS          40 TOTAL
EDM=7.63714e-07    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                INTERNAL      INTERNAL
NO.   NAME      VALUE            ERROR       STEP SIZE       VALUE
1  a1           8.94396e-03   2.69247e-02   1.96312e-04   8.94408e-03
2  a2           1.01291e-01   1.67458e-02   2.44302e-05  -1.36917e+00
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
7.251e-04 -2.218e-06
-2.218e-06  2.804e-04
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00492   1.000 -0.005
2  0.00492  -0.005  1.000
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: deactivating const optimization
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: activating const optimization
**********
**   31 **SET PRINT           1
**********
**********
**********
PARAMETER DEFINITIONS:
NO.   NAME         VALUE      STEP SIZE      LIMITS
1 a1           8.94396e-03  2.69247e-02   -1.00000e+00  1.00000e+00
2 a2           1.01291e-01  1.67458e-02    0.00000e+00  1.00000e+01
**********
**   33 **SET ERR         0.5
**********
**********
**   34 **SET PRINT           1
**********
**********
**   35 **SET STR           1
**********
NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
**********
**********
FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
FCN=118896 FROM MIGRAD    STATUS=INITIATE        8 CALLS           9 TOTAL
EDM= unknown      STRATEGY= 1      NO ERROR MATRIX
EXT PARAMETER               CURRENT GUESS       STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1           8.94396e-03   2.69247e-02   2.69290e-02   5.93783e+02
2  a2           1.01291e-01   1.67458e-02   1.67812e-02   5.89498e+02
ERR DEF= 0.5
MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=118892 FROM MIGRAD    STATUS=CONVERGED      30 CALLS          31 TOTAL
EDM=2.21121e-06    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                   STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -1.13263e-03   4.02053e-03   9.57731e-04   8.59142e-02
2  a2           9.75516e-02   2.36606e-03   5.73586e-04  -5.99867e-01
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
1.616e-05 -6.207e-08
-6.207e-08  5.598e-06
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00652   1.000 -0.007
2  0.00652  -0.007  1.000
**********
**   37 **SET ERR         0.5
**********
**********
**   38 **SET PRINT           1
**********
**********
**   39 **HESSE        1000
**********
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=118892 FROM HESSE     STATUS=OK             10 CALLS          41 TOTAL
EDM=2.21254e-06    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                INTERNAL      INTERNAL
NO.   NAME      VALUE            ERROR       STEP SIZE       VALUE
1  a1          -1.13263e-03   4.02065e-03   1.91546e-04  -1.13263e-03
2  a2           9.75516e-02   2.36614e-03   2.29434e-05  -1.37294e+00
ERR DEF= 0.5
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=0.5
1.617e-05 -9.497e-08
-9.497e-08  5.599e-06
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.00998   1.000 -0.010
2  0.00998  -0.010  1.000
[#1] INFO:Minimization -- RooAbsMinimizerFcn::setOptimizeConst: deactivating const optimization


## Chis2 fit of pdf to binned weighted dataset¶

Construct binned clone of unbinned weighted dataset

In [20]:
binnedData = wdata.binnedClone()
binnedData.Print("v")

DataStore pxData_binned (Generated From px_binned)
Contains 40 entries
Observables:
1)  x = 9.75  L(-10 - 10) B(40)  "x"
Binned Dataset pxData_binned (Generated From px_binned)
Contains 40 bins with a total weight of 43238.9
Observables:     1)  x = 9.75  L(-10 - 10) B(40)  "x"


Perform chi2 fit to binned weighted dataset using sum-of-weights errors

NB: Within the usual approximations of a chi2 fit, chi2 fit to weighted data using sum-of-weights-squared errors does give correct error estimates

In [21]:
chi2 = p2.createChi2(binnedData, ROOT.RooFit.DataError("SumW2"))
m = ROOT.RooMinimizer(chi2)
m.hesse()

Out[21]:
0
 **********
**   40 **SET PRINT           1
**********
**********
**********
PARAMETER DEFINITIONS:
NO.   NAME         VALUE      STEP SIZE      LIMITS
1 a1          -1.13263e-03  4.02065e-03   -1.00000e+00  1.00000e+00
2 a2           9.75516e-02  2.36614e-03    0.00000e+00  1.00000e+01
**********
**   42 **SET ERR           1
**********
**********
**   43 **SET PRINT           1
**********
**********
**   44 **SET STR           1
**********
NOW USING STRATEGY  1: TRY TO BALANCE SPEED AGAINST RELIABILITY
**********
**********
FIRST CALL TO USER FUNCTION AT NEW START POINT, WITH IFLAG=4.
START MIGRAD MINIMIZATION.  STRATEGY  1.  CONVERGENCE WHEN EDM .LT. 1.00e-03
FCN=32.3464 FROM MIGRAD    STATUS=INITIATE        6 CALLS           7 TOTAL
EDM= unknown      STRATEGY= 1      NO ERROR MATRIX
EXT PARAMETER               CURRENT GUESS       STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -1.13263e-03   4.02065e-03   4.02066e-03   2.74413e+01
2  a2           9.75516e-02   2.36614e-03   2.40760e-03  -2.04371e+02
MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=31.3747 FROM MIGRAD    STATUS=CONVERGED      28 CALLS          29 TOTAL
EDM=8.52339e-08    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                   STEP         FIRST
NO.   NAME      VALUE            ERROR          SIZE      DERIVATIVE
1  a1          -9.98900e-03   2.62975e-02   7.27538e-05   4.94399e-03
2  a2           1.06373e-01   1.01849e-02   2.73457e-05  -3.91519e-02
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=1
6.917e-04 -6.529e-06
-6.529e-06  1.037e-04
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.02437   1.000 -0.024
2  0.02437  -0.024  1.000
**********
**   46 **SET ERR           1
**********
**********
**   47 **SET PRINT           1
**********
**********
**   48 **HESSE        1000
**********
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=31.3747 FROM HESSE     STATUS=OK             10 CALLS          39 TOTAL
EDM=8.52317e-08    STRATEGY= 1      ERROR MATRIX ACCURATE
EXT PARAMETER                                INTERNAL      INTERNAL
NO.   NAME      VALUE            ERROR       STEP SIZE       VALUE
1  a1          -9.98900e-03   2.62976e-02   1.45508e-05  -9.98917e-03
2  a2           1.06373e-01   1.01850e-02   5.46914e-06  -1.36415e+00
EXTERNAL ERROR MATRIX.    NDIM=  25    NPAR=  2    ERR DEF=1
6.917e-04 -6.575e-06
-6.575e-06  1.037e-04
PARAMETER  CORRELATION COEFFICIENTS
NO.  GLOBAL      1      2
1  0.02455   1.000 -0.025
2  0.02455  -0.025  1.000


Plot chi^2 fit result on frame as well

In [22]:
r_chi2_wgt = m.save()
p2.plotOn(frame, LineStyle="--", LineColor="r")

Out[22]:
<cppyy.gbl.RooPlot object at 0x9133840>

## Compare fit results of chi2, L fits to (un)weighted data¶

Note that ML fit on 1Kevt of weighted data is closer to result of ML fit on 43Kevt of unweighted data than to 1Kevt of unweighted data, the reference chi^2 fit with SumW2 error gives a result closer to that of an unbinned ML fit to 1Kevt of unweighted data.

In [23]:
print("==> ML Fit results on 1K unweighted events")
r_ml_unw10.Print()
print("==> ML Fit results on 43K unweighted events")
r_ml_unw43.Print()
print("==> ML Fit results on 1K weighted events with a summed weight of 43K")
r_ml_wgt.Print()
print("==> Corrected ML Fit results on 1K weighted events with a summed weight of 43K")
r_ml_wgt_corr.Print()
print("==> Chi2 Fit results on 1K weighted events with a summed weight of 43K")
r_chi2_wgt.Print()

c = ROOT.TCanvas("rf403_weightedevts", "rf403_weightedevts", 600, 600)
frame.GetYaxis().SetTitleOffset(1.8)
frame.Draw()

c.SaveAs("rf403_weightedevts.png")

==> ML Fit results on 1K unweighted events
==> ML Fit results on 43K unweighted events
==> ML Fit results on 1K weighted events with a summed weight of 43K
==> Corrected ML Fit results on 1K weighted events with a summed weight of 43K
==> Chi2 Fit results on 1K weighted events with a summed weight of 43K

RooFitResult: minimized FCN value: 2766.49, estimated distance to minimum: 7.63714e-07
covariance matrix quality: Full, accurate covariance matrix
Status : MINIMIZE=0 HESSE=0

Floating Parameter    FinalValue +/-  Error
--------------------  --------------------------
a1    8.9440e-03 +/-  2.69e-02
a2    1.0129e-01 +/-  1.67e-02

RooFitResult: minimized FCN value: 118892, estimated distance to minimum: 2.21254e-06
covariance matrix quality: Full, accurate covariance matrix
Status : MINIMIZE=0 HESSE=0

Floating Parameter    FinalValue +/-  Error
--------------------  --------------------------
a1   -1.1326e-03 +/-  4.02e-03
a2    9.7552e-02 +/-  2.37e-03

RooFitResult: minimized FCN value: 119682, estimated distance to minimum: 1.28662e-07
covariance matrix quality: Full, accurate covariance matrix
Status : MINIMIZE=0 HESSE=0

Floating Parameter    FinalValue +/-  Error
--------------------  --------------------------
a1   -4.8560e-03 +/-  4.03e-03
a2    9.8651e-02 +/-  2.41e-03

RooFitResult: minimized FCN value: 6.84247e+06, estimated distance to minimum: 158914
covariance matrix quality: Full, accurate covariance matrix
Status : MINIMIZE=0 HESSE=0 HESSE=0

Floating Parameter    FinalValue +/-  Error
--------------------  --------------------------
a1   -4.8555e-03 +/-  3.00e-02
a2    9.8652e-02 +/-  2.99e-02

RooFitResult: minimized FCN value: 31.3747, estimated distance to minimum: 8.52317e-08
covariance matrix quality: Full, accurate covariance matrix

Floating Parameter    FinalValue +/-  Error
--------------------  --------------------------
a1   -9.9890e-03 +/-  2.63e-02
a2    1.0637e-01 +/-  1.02e-02


Info in <TCanvas::Print>: png file rf403_weightedevts.png has been created


Draw all canvases

In [24]:
from ROOT import gROOT
gROOT.GetListOfCanvases().Draw()