Opened 14 years ago

Closed 12 years ago

#194 closed feature (fixed)

Calculating a residual in NIST SANS procedures?

Reported by: flbeyer@… Owned by: srkline
Priority: major Milestone: Wish List
Component: Analysis Keywords:
Cc: Blocking:



I am trying to fit experimental data and to get some measure (other than chi) of the quality of the fit. In the Help file, calculating a residual is mentioned and the reader is referred to text "below", but nothing else on residuals is included.

Is this not possible with the SANS data analysis macros as they currently stand? Is there some other way to enable this feature? I know it is built into the Igor least squares fitting tool, for instance.

Change History (5)

comment:1 Changed 14 years ago by srkline

replied with a manual work-around... but could be added as an "advanced" panel of options

comment:2 Changed 14 years ago by srkline

  • Cc flbeyer@… removed

comment:3 Changed 12 years ago by srkline

maybe put normalized residuals (as an option) on the right axis as Jan does. Multiple data sets will be a hassle, but it's a good option to have to show the quality of fit.

comment:4 Changed 12 years ago by ajj

  • Milestone changed from Analysis Wish List to Wish List

comment:5 Changed 12 years ago by srkline

  • Resolution set to fixed
  • Status changed from new to closed

done in r743:

added two more options for curve fitting information. users can no with checkboxes add residuals to the fit

graph, and add a textbox of fit information. unchecking the boxes provides the option for removing them from the graph.

-- still need to make sure that the residuals stay in sync. if they are on the graph and then use is un-selected, the re
sidual won't update with the new fit. The textbox is fine - it's automatically killed if not selected.

Note: See TracTickets for help on using tickets.