Changes in src/sas/sasgui/perspectives/fitting/media/residuals_help.rst [99ded31:84ac3f1] in sasview
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
src/sas/sasgui/perspectives/fitting/media/residuals_help.rst
r99ded31 r84ac3f1 27 27 28 28 $\chi^2$ 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 29 an observed data set and an expected dataset (or 'theory') calculated as 33 30 34 31 .. math:: 35 32 36 \chi^2 _N37 = \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] 38 35 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 36 Fitting typically minimizes the value of $\chi^2$. For assessing the quality of 37 the model and its "fit" however, *SasView* displays the traditional reduced 38 $\chi^2_R$ which normalizes this parameter by dividing it by the number of 39 degrees of freedom (or DOF). The DOF is the number of data points being 40 considered, $N_\mathrm{pts}$, reduced by the number of free (i.e. fitted) 41 parameters, $N_\mathrm{par}$. Note that model parameters that are kept fixed do 42 *not* contribute to the DOF (they are not "free"). This reduced value is then 43 given as 42 44 43 45 .. math:: … … 47 49 / [N_\mathrm{pts} - N_\mathrm{par}] 48 50 49 The normalized $\chi^2_N$ and the reduced $\chi^2_R$ are very close to each 50 other when $N_\mathrm{pts} \gg N_\mathrm{par}$. 51 Note that this means the displayed value will vary depending on the number of 52 parameters used in the fit. In particular, when doing a calculation without a 53 fit (e.g. manually changing a parameter) the DOF will now equal $N_\mathrm{pts}$ 54 and the $\chi^2_R$ will be the smallest possible for that combination of model, 55 data set, and set of parameter values. 56 57 When $N_\mathrm{pts} \gg N_\mathrm{par}$ as it should for proper fitting, the 58 value of the reduced $\chi^2_R$ will not change very much. 51 59 52 60 For a good fit, $\chi^2_R$ tends to 1. … … 90 98 | 2015-06-08 Steve King 91 99 | 2017-09-28 Paul Kienzle 100 | 2018-03-04 Paul Butler
Note: See TracChangeset
for help on using the changeset viewer.