I have done an implementation of the Inverse Error function and the
Inverse Complementary Error function. I have called the functions
inverse_erf(z)
inverse_erfc(z)
The numerical evaluation is implemented only for real values in float
and bigfloat precision in the interval (-1,1) for the function
inverse_erf and in the interval (0,2) for the function inverse_erfc. I
have used a newton algorithm, which converges very fast.
Furthermore the derivative, the integral and a simplim%function to
handle the limits are implemented.
I would like to commit the code the next time. Do you have any comments?
Dieter Kaiser