Linear

From MantidProject

Contents 1 Summary 2 Properties 3 Description 3.1 References 4 Source Code

Summary

Performs linear least-squares regression on a spectrum (or portion of one).

Properties

Order	Name	Direction	Type	Default	Description
1	InputWorkspace	Input	Workspace	Mandatory	The name of the workspace containing the spectrum to fit.
2	OutputWorkspace	Output	Workspace1D	Mandatory	A Workspace1D containing the fit result and errors for the X values of the input spectrum.
3	SpectrumIndex	Input	integer	0	The workspace index of the spectrum to fit.
4	StartX	Input	double	Spectrum minimum	The X value to start fitting from.
5	EndX	Input	double	Spectrum maximum	The last X value to include in the fitting range.
6	FitStatus	Output	string		Empty if the fit succeeded, otherwise contains the gsl error message.
7	FitIntercept	Output	double		The intercept with the ordinate of the fitted line. c0 in the equation below.
8	FitSlope	Output	double		The slope of the fitted line. c1 in the equation below.
9	Chi2	Output	double		The χ2 value for the goodness of the fit.

Description

This algorithm fits a line, of the form Y = c0 + c1X, to the specified part of a particular spectrum. As well as outputting the result to the log (debug level) and as output properties, a 1D workspace is created with the same X bins/points as the fitted spectrum and the value and error on the fit at each point. The covariance matrix is not currently returned as an output: if this is required please contact the development team.

References

This algorithm uses the gsl linear regression algorithms gsl_fit_linear and gsl_fit_wlinear, which are documented here.

Source Code

Header Linear.h

Source Linear.cpp