F-distribution

Thread View: sci.math.num-analysis

2 messages

[T] Back to Threads [V] Tree View [G] Group Articles

2 total messages Started by maxim Fri, 08 Jan 1993 02:44

#1058

Author: maxim
Date: Fri, 08 Jan 1993 02:44

21 lines
1215 bytes

F-distribution

I am looking for either a reference or some fortran source code
that will allow me to calculate an f-value (percentile) given
that the area (probability) and degrees of freedom DF1 and DF2
are known.  There are numerous references that suggest how the
area might be calculated given that the percentile is known, but
I have yet to come across the opposite.

The algorithm must be accurate to at least the fifth decimal place.
Any suggestions would be greatly appreciated.


-- 


Paul Maxim                         | Internet:  Maxim@uwovax.uwo.ca
Department of Sociology            |  
University of Western Ontario      |
London, Ontario                    |    Phone:  519-679-2111 Ext. 5121
CANADA    N6A 5C2                  |      Fax:  519-661-3200

Re: F-distribution

#1076

Author: Herman Rubin
Date: Fri, 08 Jan 1993 17:49

25 lines
1613 bytes

In article <1993Jan7.224458.1@uwovax.uwo.ca> maxim@uwovax.uwo.ca writes:
>F-distribution

>I am looking for either a reference or some fortran source code
>that will allow me to calculate an f-value (percentile) given
>that the area (probability) and degrees of freedom DF1 and DF2
>are known.  There are numerous references that suggest how the
>area might be calculated given that the percentile is known, but
>I have yet to come across the opposite.

>The algorithm must be accurate to at least the fifth decimal place.
>Any suggestions would be greatly appreciated.

I cannot give you more than an algorithm, which I suggest should work
fairly quickly and will not require too much programming.

Since the density of the F distribution is quite simple, I suggest
that Newton's method be used to iterate.  To improve the performance,
I suggest that in the tails, the logarithm of the tail probability,
rather than the probability itself, be used.
-- 
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
Phone: (317)494-6054
hrubin@snap.stat.purdue.edu (Internet, bitnet)  
{purdue,pur-ee}!snap.stat!hrubin(UUCP)

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads