Description

Confidence interval bounds are determined for each selected parameter in turn. A given parameter's value is varied along a grid of values while the values of all the other nominally thawed parameters are allowed to float to new best-fit values (compare to covar, where the values of all the other nominally thawed parameters remain fixed to their best-fit values). This method of estimating confidence interval bounds gives truly accurate results only in special cases:

An estimated confidence interval is accurate if and only if:

• the chi^2 or logL surface in parameter space is approximately shaped like a multi-dimensional paraboloid, and
• the best-fit point is sufficiently far from parameter space boundaries.

One may determine if these conditions hold, for example, by plotting the fit statistic as a function of each parameter's values (the curve should approximate a parabola) and by examining contour plots of the fit statistics made by varying the values of two parameters at a time (the contours should be elliptical, and parameter space boundaries should be no closer than approximately 3 sigma from the best-fit point).

If no arguments are given this function, it is assumed that the data id is the default data id, and that limits should be computed for all thawed parameters.

If arguments are given, each argument is examined to see if it is a Sherpa model parameter. If model parameters are given, it is assumed that limits for only those parameters should be computed. Any argument that is not a model parameter is assumed to be a data id.