Path: utzoo!mnetor!tmsoft!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!usc!orion.oac.uci.edu!!nelson
From: nelson@ (Matt Nelson)
Newsgroups: comp.sys.handhelds
Subject: Re: MatrixWriter question
Message-ID: <276CF0BD.3492@orion.oac.uci.edu>
Date: 17 Dec 90 16:22:21 GMT
References: <1990Dec17.010220.24504@midway.uchicago.edu>
Reply-To: Matt Nelson <nelson@psroot.ps.uci.edu>
Organization: University of California, Irvine
Lines: 19
Nntp-Posting-Host: skid.ps.uci.edu

In article <1990Dec17.010220.24504@midway.uchicago.edu> keho@quads.uchicago.edu (thomas david kehoe) writes:

>Also, is there anyway to do an inverse cumulative distribution
>function?  In other words, UTPN returns the area under the
>upper tail of the normal distribution past x  (the value in
>level 1).  I often have the area under the upper tail, but
>need to find x.  No problem with the table in the back of a
>statistics book, but the 48SX can't do it..

use the solver.  suppose you had a distribution with mean=2.0 and
variance=0.5, and you wanted to find the value of x with the upper-
tail integral of 0.18.  enter the following program as EQ:

                << 2 0.5 X UTPN 0.18 - >>

go into the solver and hit (left shift)-X; you will be greeted with
"sign change" and X: 2.64726086088, which is good to 1E-12 or so.

have fun...  -matt