invnorm
|
Contents: Description, Arguments, Usage, Examples, Sub-Functions, Related Functions, Source Supported Under Version: > 1.11
Description Inverse cumulative probability function for normal distribution
Arguments and Return Values Arguments: One numeric array, with all elements between 0 and 1 (exclusive) Return Value: An array with the same organization as the input containing the value of the inverse normal function evaluated at the each element of the input. The output format is double if the input is double, float otherwise. The inverse normal function is defined such that invnorm(y) = z iff y = normcdf(-Inf, z) = normcdf(z). This function returns the z-value such that the cumulative probability in a normal distribution between negative infinity and z standard deviations from the mean equals the input. For example, if you want to find the number of standard deviations between in which lie the middle 20% of a sample (in a normal distribution), take invnorm(0.4) for the lower boundary and invnorm(0.6) for the upper boundary. This function uses a numerical approximation from Peter J. Acklam, which has a relative error of less than 1.2e-9 (in double precision). Return format is double if input is double; float otherwise. If no arguments are input, invnorm() prints its description, as shown below.
Usage Syntax: invnorm(y) 'y' - A numeric array of any format
Examples dv> invnorm() Inverse cumulative probability function for normal distribution z = invnorm(y) iff y = normcdf(z) Input must be y such that 0 < y < 1 Algorithm from http://home.online.no/~pjacklam/notes/invnorm/ Relative errors (in double precision) are less than 1.2e-9 Matches input format Works for arrays S.Marshall 02-22-2008 0 dv> invnorm(0.4//0.6) 2x1x1 array of float, bsq format [8 bytes] -0.253347 0.253347 dv> normcdf(-0.253347, 0.253347) 0.200000 dv> normcdf(-1) 0.158655 dv> invnorm(0.158655) -1.00000 |
DavinciWiki Mini-Nav Bar Contents
Contact Developers
All other topics
Recent Library Changes Created On: 02-25-2008 |
