Sigmoid Fitting Toolbox
Version 1.0 2005-Jun-25
(test line)
The following functions allow one to fit asymmetric sigmoid
functions to detection efficiency data.

  SIGMOID            Compute asymmetric sigmoid efficiency curve
  SIGMOIDFIT         Determine detection efficiencies and fit to
                     sigmoid curve
  SIGMOIDCHISQUARE   Test quality of sigmoid function fit to data

In addition, the following function is provided to demonstrate
the use of the sigmoid fitting toolbox.

  SIGMOIDEXAMPLE     Example fitting to sigmoid curve

The sigmoid curve is specified by five parameters.

  center          approximate amplitude at fifty percent
                  detection efficiency
  width           approximate width in decades
  asymmetry       dimensionless asymmetry parameter
  pedestal        minimum detection efficiency
  entablature     maximum detection efficiency

In some cases, it may be useful to explicitly force one or more of
these parameters to a specific value.  This should be done in the
function SIGMOID.  Typical constraints are a pedestal of zero, an
entablature of unit, and an asymmetry of zero.

Note that the center parameter is not the same as the amplitude at
50 percent detection efficiency unless the pedestal is zero, the
entablature is unity, and the asymmetry is zero.  In general, the
amplitude at 50 percent deteection efficiency can be determined by
interpolation from the resulting sigmoid curve.  Perhaps there is an
analytical expression for it, but it is has not been investigated.

Note that the sigmoid fitting toolbox requires the Matlab Statistics
toolbox.

Shourov K. Chatterji
shourov@ligo.mit.edu

Lindy L. Blackburn
lindy@ligo.mit.edu

2005-Jun-25

$Id: README.html 6645 2010-12-09 16:09:14Z diego.menendez@LIGO.ORG $