Support Vector Regression based insulin dose forecasting for type II diabetes patients
In this paper, we compared the performance of Support Vector Regression (SVR), based insulin dose forecasting for type II diabetes patients, with Artificial Neural Network (ANN) learning with BackPropagation, Conjugate Gradient Descent, Levenberg-Marquardt, Quasi Newton and Quick Propagation respect...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Conference Proceeding |
| Published: |
2018
|
| Subjects: | |
| Online Access: | https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84906998294&origin=inward http://cmuir.cmu.ac.th/jspui/handle/6653943832/60439 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Chiang Mai University |
| Summary: | In this paper, we compared the performance of Support Vector Regression (SVR), based insulin dose forecasting for type II diabetes patients, with Artificial Neural Network (ANN) learning with BackPropagation, Conjugate Gradient Descent, Levenberg-Marquardt, Quasi Newton and Quick Propagation respectively. The methodology of this study started from collecting data of diabetes patients from Chiang Mai Maharaj Hospital. A series of experiments have been conducted for six approaches. After the learning processes had been accomplished, a performance of SVR was compared with the others in term of mean absolute deviation (MAD). The experimental results suggest that SVR can be dramatically trained in a shorter time than the others. In addition, insulin dose level, calculated by all of six approaches close to insulin level, was controlled by doctor. Moreover, SVR also provided the most robust among these five approaches of ANN. © 2008 ICQR. |
|---|