@cindex error function
@cindex erf(x)
@cindex erfc(x)

The error function is described in Abramowitz & Stegun, Chapter 7.  The
functions in this section are declared in the header file
@file{gsl_sf_erf.h}.

@menu
* Error Function::              
* Complementary Error Function::  
* Log Complementary Error Function::  
* Probability functions::       
@end menu

@node Error Function
@subsection Error Function

@deftypefun double gsl_sf_erf (double @var{x})
@deftypefunx int gsl_sf_erf_e (double @var{x}, gsl_sf_result * @var{result})
These routines compute the error function @c{$\erf(x)$}
@math{erf(x)}, where
@c{$\erf(x) = (2/\sqrt{\pi}) \int_0^x dt \exp(-t^2)$}
@math{erf(x) = (2/\sqrt(\pi)) \int_0^x dt \exp(-t^2)}.
@comment Exceptional Return Values: none
@end deftypefun

@node Complementary Error Function
@subsection Complementary Error Function

@deftypefun double gsl_sf_erfc (double @var{x})
@deftypefunx int gsl_sf_erfc_e (double @var{x}, gsl_sf_result * @var{result})
These routines compute the complementary error function 
@c{$\erfc(x) = 1 - \erf(x) = (2/\sqrt{\pi}) \int_x^\infty \exp(-t^2)$}
@math{erfc(x) = 1 - erf(x) = (2/\sqrt(\pi)) \int_x^\infty \exp(-t^2)}.
@comment Exceptional Return Values: none
@end deftypefun


@node Log Complementary Error Function
@subsection Log Complementary Error Function

@deftypefun double gsl_sf_log_erfc (double @var{x})
@deftypefunx int gsl_sf_log_erfc_e (double @var{x}, gsl_sf_result * @var{result})
These routines compute the logarithm of the complementary error function
@math{\log(\erfc(x))}.
@comment Exceptional Return Values: none
@end deftypefun


@node Probability functions
@subsection Probability functions

The probability functions for the Normal or Gaussian distribution are
described in Abramowitz & Stegun, Section 26.2.

@deftypefun double gsl_sf_erf_Z (double @var{x})
@deftypefunx int gsl_sf_erf_Z_e (double @var{x}, gsl_sf_result * @var{result})
These routines compute the Gaussian probability density function 
@c{$Z(x) = (1/\sqrt{2\pi}) \exp(-x^2/2)$} 
@math{Z(x) = (1/\sqrt@{2\pi@}) \exp(-x^2/2)}.  
@end deftypefun

@deftypefun double gsl_sf_erf_Q (double @var{x})
@deftypefunx int gsl_sf_erf_Q_e (double @var{x}, gsl_sf_result * @var{result})
These routines compute the upper tail of the Gaussian probability
function 
@c{$Q(x) = (1/\sqrt{2\pi}) \int_x^\infty dt \exp(-t^2/2)$}
@math{Q(x) = (1/\sqrt@{2\pi@}) \int_x^\infty dt \exp(-t^2/2)}.
@comment Exceptional Return Values: none
@end deftypefun

@cindex hazard function, normal distribution
@cindex Mill's ratio, inverse
The @dfn{hazard function} for the normal distribution, 
also known as the inverse Mill's ratio, is defined as,
@tex
\beforedisplay
$$
h(x) = {Z(x)\over Q(x)} = \sqrt{2 \over \pi} {\exp(-x^2 / 2) \over \erfc(x/\sqrt 2)}
$$
\afterdisplay
@end tex
@ifinfo

@example
h(x) = Z(x)/Q(x) = \sqrt@{2/\pi@} \exp(-x^2 / 2) / \erfc(x/\sqrt 2)
@end example

@end ifinfo
@noindent
It decreases rapidly as @math{x} approaches @math{-\infty} and asymptotes
to @math{h(x) \sim x} as @math{x} approaches @math{+\infty}.

@deftypefun double gsl_sf_hazard (double @var{x})
@deftypefunx int gsl_sf_hazard_e (double @var{x}, gsl_sf_result * @var{result})
These routines compute the hazard function for the normal distribution.
@comment Exceptional Return Values: GSL_EUNDRFLW
@end deftypefun