Changeset a7c6f38 in sasview


Ignore:
Timestamp:
Mar 4, 2018 6:07:55 PM (4 years ago)
Author:
butler
Branches:
master, magnetic_scatt, release-4.2.2, ticket-1009, ticket-1094-headless, ticket-1242-2d-resolution, ticket-1243, ticket-1249, unittest-saveload
Children:
84ac3f1
Parents:
943e17a
Message:

Clarify use of chi2 in GUI and documentation

Changed Gui name and tooltip to reflect we are using reduced chi2 not
simply Chi
2/Npts. Also edited the documenataion to finish making that
clear and in the process cleaned up a few old (i.e. incorrect)
statements.

Location:
src/sas/sasgui/perspectives/fitting
Files:
3 edited

Legend:

Unmodified
Added
Removed
  • src/sas/sasgui/perspectives/fitting/fitpage.py

    rbfeb823 ra7c6f38  
    365365        # StaticText for chi2, N(for fitting), Npts + Log/linear spacing 
    366366        self.tcChi = BGTextCtrl(self, wx.ID_ANY, "-", size=(75, 20), style=0) 
    367         self.tcChi.SetToolTipString("Chi2/Npts(Fit)") 
     367        self.tcChi.SetToolTipString("Chi2/DOF (DOF=Npts-Npar fitted)") 
    368368        self.Npts_fit = BGTextCtrl(self, wx.ID_ANY, "-", size=(75, 20), style=0) 
    369369        self.Npts_fit.SetToolTipString( 
     
    391391        self.points_sizer.Add(self.pointsbox) 
    392392 
    393         box_description_1 = wx.StaticText(self, wx.ID_ANY, '   Chi2/Npts') 
     393        box_description_1 = wx.StaticText(self, wx.ID_ANY, 'Reduced Chi2') 
    394394        box_description_2 = wx.StaticText(self, wx.ID_ANY, 'Npts(Fit)') 
    395395 
  • src/sas/sasgui/perspectives/fitting/media/fitting_help.rst

    r5005ae0 ra7c6f38  
    426426See :ref:`Assessing_Fit_Quality`. 
    427427 
    428 The objective of model-fitting is to find a *physically-plausible* model, and set 
    429 of model parameters, that generate a theory that reproduces the experimental data 
    430 and gives residual values as close to zero as possible. 
     428The objective of model-fitting is to find a *physically-plausible* model, and 
     429set of model parameters, that generate a theory that reproduces the experimental 
     430data and minimizes the values of the residuals. 
    431431 
    432432Change the default values of the model parameters by hand until the theory line 
    433 starts to represent the experimental data. Then uncheck the tick boxes alongside 
    434 all parameters *except* the 'background' and the 'scale'. Click the *Fit* button. 
    435 SasView will optimise the values of the 'background' and 'scale' and also display 
    436 the corresponding uncertainties on the optimised values. 
    437  
    438 *NB: If no uncertainty is shown it generally means that the model is not very* 
    439 *dependent on the corresponding parameter (or that one or more parameters are* 
    440 *'correlated').* 
    441  
    442 In the bottom left corner of the *Fit Page* is a box displaying the normalised value 
    443 of the statistical $\chi^2$ parameter returned by the optimiser. 
     433starts to represent the experimental data. Then check the tick boxes alongside 
     434the 'background' and the 'scale' parameters. Click the *Fit* button. SasView 
     435will optimise the values of the 'background' and 'scale' and also display the 
     436corresponding uncertainties on the optimised values. 
     437 
     438.. note:: 
     439   If a parameter uncertainty is unrealistically large, or if it displays as 
     440   NaN either the model is a poor representation of the data, the parameter in 
     441   question is either highly correlated with one or more other fitted parameters 
     442   or at least that the model is relatively insensitive to the value of that 
     443   particular parameter. 
     444 
     445In the bottom left corner of the *Fit Page* is a box displaying the normalised 
     446value of the statistical $\chi^2$ parameter returned by the optimiser. 
    444447 
    445448Now check the box for another model parameter and click *Fit* again. Repeat this 
    446449process until most or all parameters are checked and have been optimised. As the 
    447 fit of the theory to the experimental data improves the value of 'chi2/Npts' will 
    448 decrease. A good model fit should easily produce values of 'chi2/Npts' that are 
     450fit of the theory to the experimental data improves, the value of 'Reduced Chi2' 
     451will decrease. A good model fit should produce values of Reduced Chi2 close 
    449452close to one, and certainly <100. See :ref:`Assessing_Fit_Quality`. 
    450453 
     
    512515*FitPage*'s. 
    513516 
    514 Note that the chi2/Npts value returned is the SUM of the chi2/Npts of each fit. To 
    515 see the chi2/Npts value for a specific *FitPage*, click the *Compute* button at the 
    516 bottom of that *FitPage* to recalculate. Also see :ref:`Assessing_Fit_Quality`. 
     517Note that the Reduced Chi2 value returned is the SUM of the Reduced Chi2 of 
     518each fit. To see the Reduced Chi2 value for a specific *FitPage*, click the  
     519*Compute* button at the bottom of that *FitPage* to recalculate. Note that in 
     520doing so the degrees of freedome will be set to Npts. 
     521See :ref:`Assessing_Fit_Quality`. 
    517522 
    518523Simultaneous Fits with Constraints 
     
    538543*FitPage*'s. 
    539544 
    540 Note that the chi2/Npts value returned is the SUM of the chi2/Npts of each fit. To 
    541 see the chi2/Npts value for a specific *FitPage*, click the *Compute* button at the 
    542 bottom of that *FitPage* to recalculate. Also see :ref:`Assessing_Fit_Quality`. 
    543  
     545Note that the Reduced Chi2 value returned is the SUM of the Reduced Chi2 of 
     546each fit. To see the Reduced Chi2 value for a specific *FitPage*, click the  
     547*Compute* button at the bottom of that *FitPage* to recalculate. Note that in 
     548doing so the degrees of freedome will be set to Npts. 
     549See :ref:`Assessing_Fit_Quality`.  Moreover in the case of constraints the 
     550degrees of freedom are less than one might think do to those constraints. 
    544551.. ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ 
    545552 
  • src/sas/sasgui/perspectives/fitting/media/residuals_help.rst

    r99ded31 ra7c6f38  
    2727 
    2828$\chi^2$ is a statistical parameter that quantifies the differences between 
    29 an observed data set and an expected dataset (or 'theory'). 
    30  
    31 When showing the a model with the data, *SasView* displays this parameter 
    32 normalized to the number of data points, $N_\mathrm{pts}$ such that 
     29an observed data set and an expected dataset (or 'theory') calculated as 
    3330 
    3431.. math:: 
    3532 
    36   \chi^2_N 
    37   =  \sum[(Y_i - \mathrm{theory}_i)^2 / \mathrm{error}_i^2] / N_\mathrm{pts} 
     33  \chi^2 
     34  =  \sum[(Y_i - \mathrm{theory}_i)^2 / \mathrm{error}_i^2] 
    3835 
    39 When performing a fit, *SasView* instead displays the reduced $\chi^2_R$, 
    40 which takes into account the number of fitting parameters $N_\mathrm{par}$ 
    41 (to calculate the number of 'degrees of freedom'). This is computed as 
     36Fitting typically minimizes the value of $\chi^2$.  However, for assessing the 
     37quality of the model and its "fit" this parameter is not terribly helpful on its 
     38own.  Thus *SasView* instead displays a normalized version of this parameter, 
     39using the traditional reduced $\chi^2_R$.  This is the $\chi^2$ divided by the 
     40degrees of freedom (or DOF). The DOF is simply the number of data points being 
     41considered reduced by the number of free (i.e. fitted) parameters. Note that 
     42model parameters that are kept fixed do *not* contribute to the DOF (they are 
     43not"free". This reduced value is then given as 
    4244 
    4345.. math:: 
     
    4749  / [N_\mathrm{pts} - N_\mathrm{par}] 
    4850 
    49 The normalized $\chi^2_N$ and the reduced $\chi^2_R$ are very close to each 
    50 other when $N_\mathrm{pts} \gg N_\mathrm{par}$. 
     51where $N_\mathrm{par}$ is the number of *fitted* parameters. Note that this 
     52means the displayed value will vary depending on the number of parameters used 
     53in the fit.  In particular, when doing a calculation without a fit (e.g. 
     54manually changing a parameter) the DOF will now equal $N_\mathrm{pts}$ and the 
     55$\chi^2_R$ will be the smallest possible for that combination of model, data 
     56set and set of parameter values. 
     57 
     58When $N_\mathrm{pts} \gg N_\mathrm{par}$ as it should for proper fitting, the 
     59value of the reduced $\chi^2_R$ will not change very much. 
    5160 
    5261For a good fit, $\chi^2_R$ tends to 1. 
Note: See TracChangeset for help on using the changeset viewer.