Date: Mon, 9 May 2005 12:06:58 -0600
Reply-To: William Kossack <kossackw@NJC.ORG>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: William Kossack <kossackw@NJC.ORG>
Organization: Posted via Supernews, http://www.supernews.com
Subject: Re: computing area under a curve
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
A couple people have pointed me to a sugi paper on using the trapizoid
rule to computer AUC.
This will be helpful. They are currently doing something similar in
excel to compute AUC.
A more critical problem is that the response curves appear to have 3
different shapes. ie bell shape or 'normal', ski slope, and irregular
(for example the value goes up then down then up again and finally down).
The question is can you really use AUC to compare curves of such
Paige Miller wrote:
> William Kossack wrote:
>> We have a project that is measuring values over time for a
>> biochemical response to a stimulus. In a 'normal' response, the
>> curve of the response goes from a baseline value then increases to a
>> point then decreases.
>> For our analysis we need to compute the area under the curve created
>> by values measured at a regular interval. Each experiment results in
>> 3-5 values taken at 15 minute intervals.
>> Does SAS have a function that can compute the area under a curve? We
>> are looking for something that can estimate the curve thus producing
>> maybe more accurate area under the curve. We are also looking for a
>> way to compare several curve shapes we see in the data.
> As far as I know, SAS does not have a built-in "area under the curve"
> function. You have to write your own (or borrow from someone who has
> already done it). The trapezoid rule is not hard to program.