Refinement: adjusting improper region bounds
There may be several cases, when HyperLab's default region bound determination is considered sub-optimal by the spectroscopist.
The figure depicts a well-fitted peak, but HyperLab did not notice the change in background type after channel 305. The residual and the high Chi-squared value (3.1) designates the possible problem with the fit.
If you take a closer look at the background, you can observe the sharp changing at the right side. The modification of the background boundary is sugges­ted in this case, so let us try to move the right region boundary leftwards.
Click on the blue arrow at the right boundary of the fitted region, and drag it to its new position, e.g. to channel 305.
When you release the mouse button, HyperLab modifies the fit model, performs a refit and displays the new region.
The region's right bound has changed, and also a different background function is used. The improved Chi-squared value shows the better fit.
If you look at the left side of the region, you will notice that the same problem of changing background may occur there, so it is worth a try moving the left region limit to channel 288.
After that change, the Chi-squared may be even lower, and non-random tendencies will mostly disappear from the residual.
If you are not satisfied with the results, just click the Undo button, and your previous fit will be restored.
Copyright © 1998-2007 by HyperLabs Software Budapest, Hungary