Page History

Now that we have learned how to create functions and to configure the fitter we are ready to start fitting. The following code shows how to perform a simple chi squared fit to a one dimensional histogram:

// The AIDA Factories
IAnalysisFactory af = IAnalysisFactory.create();
ITree = af.createTreeFactory().create();
IFunctionFactory ff = af.createFunctionFactory(tree);
IFitter fitter = af.createFitFactory().createFitter("chi2");

// Fit a gaussian to a given histogram h1
IFitResult result = fitter.fit(h1,"g");

Cloning When Fitting

When a function is provided to the fitter the user has to be aware that the AIDA prescription is to leave the input function unchanged, create a clone, perfom the fit on the clone and return the clone in the fit result. There advantage to this approach is that the input function can be reused as is for multiple identical fits while the fitted-cloned function is left unchanged in the fit result.
To override the above behavior the user can create a fitter by setting the noClone option to true in the options:

// Create a fitter that does not clone the input function
IFitter fitter = af.createFitFactory().createFitter("chi2","minuit","noClone=true");

Fit Result

The outcome of a fit is an IFitResult. It contains the fitted function, the error on the parameters, the covariance matrix etc.
The advantage of putting the result of the fit into a result object is that additional fits, error analysis, contours or scans can be easily calculated by using the information stored in it.