|
Abstract
|
|
|---|---|
| This work presents the influence of some factors on the precision of a glucose predictor. The predictor, based on neural networks (NNM) was individually trained with continuous glucose data from nine patients. The precision was calculated as the root mean square error (RMSE) between the original and predicted profiles (Pérez-Gandía et al, 2008). We analyze the reasons of the inter-patient precision variability and the optimum training-set size to improve the model performance. | |
|
International
|
Si |
|
Congress
|
2nd Conference on Advanced Technologies & Treatments for Diabetes. Atenas, A156, 2009. |
|
|
960 |
|
Place
|
Atenas |
|
Reviewers
|
Si |
|
ISBN/ISSN
|
00000000 |
|
|
|
|
Start Date
|
25/02/2009 |
|
End Date
|
27/02/2009 |
|
From page
|
156 |
|
To page
|
156 |
|
|
2nd Conference on Advanced Technologies & Treatments for Diabetes |