Go back to the model editing tool, pull down the Tools menu and select the Least-Squares Tool command. This will start the least-squares tool shown in Figure 2.6. Move this to a place on your screen where you can see the least-squares tool, the model editing tool and at least one baseline tool.
Figure 2.6: The Least-Squares Tool.
All you need to do now is to pull down the Fitter menu and click on the Fit Data command. You will see a working dialogue that tells you that the model is being optimized and values will appear in the fields to the upper right of the tool. The model shown in the model editor and the model visibilities will be updated every time the least-squares tool completes an iteration. Some of the changes may be too small for you to see them, however.
Each iteration may take several minutes. This time will increase with the number of data points and the number of model parameters. The application will be comatose durinf each iteration.
If all goes well, you will eventually see a display like that in
Figure 2.7. This shows that the least-squares fit has
converged on a good solution. You will see the final parameters of
the component parameters together with their estimated standard errors
and the final value of the statistic divided by the number of
datapoints. If the
value per point is about one, you have an
exceptionally good fit with the data deviating from the model
visibilities by about one standard error on average. Realistically,
a
value around 10 or less indicates a pretty good fit.
Figure 2.7: The Results of a Good Least-Squares Fit.
It is unlikely that the fitting algorithm will fail with such a simpel model. If it does, consult the reference manual for an explanation of the failure.
If you don't like the model that the least-squares tool gives you, you can use the Undo command in the model editing tool to reset the model back to the way it was before you attempted the fit.