syntax for AUC
Locked
 
|
Hi all
Does anybody have syntax to calculate the positive incremental area under the curve (or the area under the curve above the baseline)? The only thing I could find with Google is SAS syntax. Thanks! Cathie |
|
|
Here is some AUC and IAUC (incremental, which adjusts for baseline)
syntax for a 2 hour oral glucose tolerance test. Should get you on your way. Good luck! compute gl_auc = sum( (.25*mean(glun5a, glu10a)), (.167*mean(glu10a, glu20a)), (.167*mean(glu20a, glu30a)), (.167*mean(glu30a, glu40a)), (.167*mean(glu40a, glu50a)), (.167*mean(glu50a, glu60a)), (.167*mean(glu60a, glu70a)), (.167*mean(glu70a, glu80a)), (.167*mean(glu80a, glu90a)), (.167*mean(glu90a, glu100a)), (.167*mean(glu100a, glu110a)), (.167*mean(glu110a, glu120a)), (.167*mean(glu120a, glu130a)), (.167*mean(glu130a, glu140a)), (.167*mean(glu140a, glu150a)), (.167*mean(glu150a, glu160a)), (.167*mean(glu160a, glu170a)), (.167*mean(glu170a, glu180a))). compute gl_iauc = sum( ((.25*mean(glun5a, glu10a))-(glun5a*.25)), ((.167*mean(glu10a, glu20a))-(glun5a*.167)), ((.167*mean(glu20a, glu30a))-(glun5a*.167)), ((.167*mean(glu30a, glu40a))-(glun5a*.167)), ((.167*mean(glu40a, glu50a))-(glun5a*.167)), ((.167*mean(glu50a, glu60a))-(glun5a*.167)), ((.167*mean(glu60a, glu70a))-(glun5a*.167)), ((.167*mean(glu70a, glu80a))-(glun5a*.167)), ((.167*mean(glu80a, glu90a))-(glun5a*.167)), ((.167*mean(glu90a, glu100a))-(glun5a*.167)), ((.167*mean(glu100a, glu110a))-(glun5a*.167)), ((.167*mean(glu110a, glu120a))-(glun5a*.167)), ((.167*mean(glu120a, glu130a))-(glun5a*.167)), ((.167*mean(glu130a, glu140a))-(glun5a*.167)), ((.167*mean(glu140a, glu150a))-(glun5a*.167)), ((.167*mean(glu150a, glu160a))-(glun5a*.167)), ((.167*mean(glu160a, glu170a))-(glun5a*.167)), ((.167*mean(glu170a, glu180a))-(glun5a*.167))). exe. Christi On Apr 14, 2010, at 8:00 AM, CATHIE ATKINSON wrote: > Hi all > > Does anybody have syntax to calculate the positive incremental area > under the curve (or the area under the curve above the baseline)? > The only thing I could find with Google is SAS syntax. > > Thanks! > Cathie ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
Hi Cathie,
In addition to Christianne's examples, there are more general methods for both total and incremental AUC at: http://www.spsstools.net/Syntax/AUC/AreaUnderTheCurveUsingTrapezoidalIntegration.txt http://www.spsstools.net/Syntax/AUC/IncrementalAUC.txt Cheers, Kylie. Quoting Christianne Joy Lane <[hidden email]>: > Here is some AUC and IAUC (incremental, which adjusts for baseline) > syntax for a 2 hour oral glucose tolerance test. Should get you on > your way. Good luck! > > compute gl_auc = sum( (.25*mean(glun5a, glu10a)), > (.167*mean(glu10a, glu20a)), > (.167*mean(glu20a, glu30a)), > (.167*mean(glu30a, glu40a)), > (.167*mean(glu40a, glu50a)), > (.167*mean(glu50a, glu60a)), > (.167*mean(glu60a, glu70a)), > (.167*mean(glu70a, glu80a)), > (.167*mean(glu80a, glu90a)), > (.167*mean(glu90a, glu100a)), > (.167*mean(glu100a, glu110a)), > (.167*mean(glu110a, glu120a)), > (.167*mean(glu120a, glu130a)), > (.167*mean(glu130a, glu140a)), > (.167*mean(glu140a, glu150a)), > (.167*mean(glu150a, glu160a)), > (.167*mean(glu160a, glu170a)), > (.167*mean(glu170a, glu180a))). > > compute gl_iauc = sum( ((.25*mean(glun5a, glu10a))-(glun5a*.25)), > ((.167*mean(glu10a, glu20a))-(glun5a*.167)), > ((.167*mean(glu20a, glu30a))-(glun5a*.167)), > ((.167*mean(glu30a, glu40a))-(glun5a*.167)), > ((.167*mean(glu40a, glu50a))-(glun5a*.167)), > ((.167*mean(glu50a, glu60a))-(glun5a*.167)), > ((.167*mean(glu60a, glu70a))-(glun5a*.167)), > ((.167*mean(glu70a, glu80a))-(glun5a*.167)), > ((.167*mean(glu80a, glu90a))-(glun5a*.167)), > ((.167*mean(glu90a, glu100a))-(glun5a*.167)), > ((.167*mean(glu100a, glu110a))-(glun5a*.167)), > ((.167*mean(glu110a, glu120a))-(glun5a*.167)), > ((.167*mean(glu120a, glu130a))-(glun5a*.167)), > ((.167*mean(glu130a, glu140a))-(glun5a*.167)), > ((.167*mean(glu140a, glu150a))-(glun5a*.167)), > ((.167*mean(glu150a, glu160a))-(glun5a*.167)), > ((.167*mean(glu160a, glu170a))-(glun5a*.167)), > ((.167*mean(glu170a, glu180a))-(glun5a*.167))). > > exe. > > Christi > > On Apr 14, 2010, at 8:00 AM, CATHIE ATKINSON wrote: > > > Hi all > > > > Does anybody have syntax to calculate the positive incremental area > > under the curve (or the area under the curve above the baseline)? > > The only thing I could find with Google is SAS syntax. > > > > Thanks! > > Cathie > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
thanks Christianne and Kylie. I do already have syntax to do the incremental AUC but I'm hoping something exists to calculate the positive incremental, that is just the portion above the baseline. I'm looking at salivary cortisol which may go up and down during the experimental period. So where the incremental AUC would be negative and all the data points are below the baseline, the positive incremental AUC would be 0. It's probably straightforward if there's only 1 segment above baseline but where there are multiple segments, I'm not sure how to handle that. I haven't tried to work it out myself yet in the hopes that it exists somewhere. Actually, as I'm writing this I'm getting ideas.....
----- Original Message ----- From: Kylie Lange <[hidden email]> Date: Wednesday, April 14, 2010 7:44 pm Subject: Re: syntax for AUC To: [hidden email] > Hi Cathie, > > In addition to Christianne's examples, there are more general > methods for both > total and incremental AUC at: > > http://www.spsstools.net/Syntax/AUC/AreaUnderTheCurveUsingTrapezoidalIntegration.txt > > http://www.spsstools.net/Syntax/AUC/IncrementalAUC.txt > > Cheers, > Kylie. > > > Quoting Christianne Joy Lane <[hidden email]>: > > > Here is some AUC and IAUC (incremental, which adjusts for baseline) > > syntax for a 2 hour oral glucose tolerance test. Should get > you on > > your way. Good luck! > > > > compute gl_auc = sum( (.25*mean(glun5a, glu10a)), > > (.167*mean(glu10a, glu20a)), > > (.167*mean(glu20a, glu30a)), > > (.167*mean(glu30a, glu40a)), > > (.167*mean(glu40a, glu50a)), > > (.167*mean(glu50a, glu60a)), > > (.167*mean(glu60a, glu70a)), > > (.167*mean(glu70a, glu80a)), > > (.167*mean(glu80a, glu90a)), > > (.167*mean(glu90a, glu100a)), > > (.167*mean(glu100a, glu110a)), > > (.167*mean(glu110a, glu120a)), > > (.167*mean(glu120a, glu130a)), > > (.167*mean(glu130a, glu140a)), > > (.167*mean(glu140a, glu150a)), > > (.167*mean(glu150a, glu160a)), > > (.167*mean(glu160a, glu170a)), > > (.167*mean(glu170a, glu180a))). > > > > compute gl_iauc = sum( ((.25*mean(glun5a, glu10a))-(glun5a*.25)), > > ((.167*mean(glu10a, glu20a))-(glun5a*.167)), > > ((.167*mean(glu20a, glu30a))-(glun5a*.167)), > > ((.167*mean(glu30a, glu40a))-(glun5a*.167)), > > ((.167*mean(glu40a, glu50a))-(glun5a*.167)), > > ((.167*mean(glu50a, glu60a))-(glun5a*.167)), > > ((.167*mean(glu60a, glu70a))-(glun5a*.167)), > > ((.167*mean(glu70a, glu80a))-(glun5a*.167)), > > ((.167*mean(glu80a, glu90a))-(glun5a*.167)), > > ((.167*mean(glu90a, glu100a))-(glun5a*.167)), > > ((.167*mean(glu100a, glu110a))-(glun5a*.167)), > > ((.167*mean(glu110a, glu120a))-(glun5a*.167)), > > ((.167*mean(glu120a, glu130a))-(glun5a*.167)), > > ((.167*mean(glu130a, glu140a))-(glun5a*.167)), > > ((.167*mean(glu140a, glu150a))-(glun5a*.167)), > > ((.167*mean(glu150a, glu160a))-(glun5a*.167)), > > ((.167*mean(glu160a, glu170a))-(glun5a*.167)), > > ((.167*mean(glu170a, glu180a))-(glun5a*.167))). > > > > exe. > > > > Christi > > > > On Apr 14, 2010, at 8:00 AM, CATHIE ATKINSON wrote: > > > > > Hi all > > > > > > Does anybody have syntax to calculate the positive > incremental area > > > under the curve (or the area under the curve above the baseline)? > > > The only thing I could find with Google is SAS syntax. > > > > > > Thanks! > > > Cathie > > > > ===================== > > To manage your subscription to SPSSX-L, send a message to > > [hidden email] (not to SPSSX-L), with no body text > except the > > command. To leave the list, send the command > > SIGNOFF SPSSX-L > > For a list of commands to manage subscriptions, send the command > > INFO REFCARD > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text > except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD > > > -- > BEGIN-ANTISPAM-VOTING-LINKS > ------------------------------------------------------ > > Teach CanIt if this mail (ID 1025910580) is spam: > Spam: > https://antispam.osu.edu/b.php?i=1025910580&m=1e5600261eeb&c=sNot spam: https://antispam.osu.edu/b.php?i=1025910580&m=1e5600261eeb&c=n > Forget vote: > https://antispam.osu.edu/b.php?i=1025910580&m=1e5600261eeb&c=f--- > --------------------------------------------------- > END-ANTISPAM-VOTING-LINKS > |
|
|
Hi Cathie, I’m haven’t heard the terminology “positive
incremental” before – in my experience simply “incremental”
is used to mean only the above baseline segments. So unless I’m missing
something, then the second syntax program I linked to below (incremental AUC) should
do what you want. It ignores all segments below the baseline value and a case
with all values lower than baseline will return an iauc2 value of 0. I believe
this is what you are after? Do you have examples of cases where it doesn’t
return the expected result of 0? Cheers, Kylie. From: CATHIE ATKINSON [mailto:[hidden email]]
thanks Christianne and Kylie. I do already have syntax
to do the incremental AUC but I'm hoping something exists to calculate the
positive incremental, that is just the portion above the baseline. I'm
looking at salivary cortisol which may go up and down during the experimental
period. So where the incremental AUC would be negative and all the data
points are below the baseline, the positive incremental AUC would be 0.
It's probably straightforward if there's only 1 segment above baseline but
where there are multiple segments, I'm not sure how to handle that. I
haven't tried to work it out myself yet in the hopes that it exists
somewhere. Actually, as I'm writing this I'm getting ideas..... |
|
|
In reply to this post by Cathie Atkinson
My apologies! I have to admit I glanced at it quickly and didn't run it before I responded. I'm thrilled there is a solution and I'm off to play with data. Thanks again!
----- Original Message ----- From: Kylie Lange <[hidden email]> Date: Thursday, April 15, 2010 8:25 pm Subject: RE: syntax for AUC To: 'CATHIE ATKINSON' <[hidden email]> Cc: [hidden email]
|
| Free forum by Nabble | Edit this page |