next up previous contents
Next: Beta profiles Up: The procedure Previous: The procedure   Contents


Constraining the fit

The progress of obtaining best fits can be manipulated by applying restricting conditions to the procedure. This is done by indicating parameters for the linear condition

\begin{displaymath}
\sum_{i=0}^{nlines}f_i\times a_i=v
\end{displaymath} (2)

with a weight factor $f_i$ for each line and a right hand side $v$. When all weight factors are zero, no restricting conditions will be applied. Only lines with weight factors $f_j\ne0$ will have to meet the restrictions during the iteration, all lines with $f_j=0$ are free, will have to adapt, though, to the restrictions of the other lines. For example, with Eq. 2 an expected count number for a line can be set fixed, by setting the corresponding weight factor $f_i=1$ and setting $v$ to the expected line counts. All other lines will then have to adapt to the one line having the desired line counts. Also line ratios can be fixed, by setting, e.g., $f_i=0.5$, $f_j=-1.$, and $v=0$. The ratio of $a_j/a_i$ will then be 0.5 at the end of the fit, and the rest will end up with the best fit under this restriction.


next up previous contents
Next: Beta profiles Up: The procedure Previous: The procedure   Contents
Jan-Uwe Ness
2003-05-23