|
Abstract
|
|
|---|---|
| The performance of glucose predictors from continuous glucose monitoring devices is usually assessed by considering two key aspects: the accuracy, usually calculated as the Root Mean Square Error (RMSE) between the original and predicted CGM profiles, and the Mean Delay (MD), usually calculated as the relative delay in crossing some fixed thresholds. However, these parameters are not sufficient for assess the model utility in extreme hypo/hyper ranges and when the glucose trend changes. We present additional evaluation parameters to better characterize the predictor behavior in these extreme situations. | |
|
International
|
No |
|
Congress
|
2nd Conference on Advanced Technologies & Treatments for Diabetes |
|
|
960 |
|
Place
|
Athens |
|
Reviewers
|
Si |
|
ISBN/ISSN
|
00000000 |
|
|
|
|
Start Date
|
25/02/2009 |
|
End Date
|
28/02/2009 |
|
From page
|
1 |
|
To page
|
1 |
|
|
2nd Conference on Advanced Technologies & Treatments for Diabetes, libro de actas |