|
Activity Number:
|
568
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305416 |
|
Title:
|
Free-Living Inferential Glucose Dynamic Modeling Using Only Noninvasive Input
|
|
Author(s):
|
Derrick Rollins*+ and Lucas Beverlin and Nisarg Vyas and Gregory Welk and Warren Franke
|
|
Companies:
|
Iowa State University and Iowa State University and BodyMedia, Inc. and Iowa State University and Iowa State University
|
|
Address:
|
, , ,
|
|
Keywords:
|
|
|
Abstract:
|
The goal of this work is the development of a causation modeling methodology to accurately infer plasma glucose over an extended period of time from non-invasive input variables. Data for subjects were collected under free-living conditions to extensively cover the input space representing the normal behavior. Models were accurately identified and tested using eleven food and activity variables. Training consisted of 3 days to 3 weeks and testing data covered one day to 3 weeks. Results are presented for Types 1 and 2 diabetic subjects. The best case gave a training R-sq of 0.92 and a testing fitted correlation coefficient of 0.96. For test data, over a three week period, the average absolute error (AAE) was as low as 10.0 mg/dL and significantly better than the glucose lancet meter AAE for replicated measurements of 15.3 mg/dL.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
