|
Activity Number:
|
369
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #301957 |
|
Title:
|
An Artificial Immune Network Based Classification Approach to ECG Monitoring Applications
|
|
Author(s):
|
Honggang Wang and Hua Fang*+
|
|
Companies:
|
University of Nebraska-Lincoln and University of Nebraska-Lincoln
|
|
Address:
|
2222 Vine St. Apt. 307, Lincoln, NE, 68503,
|
|
Keywords:
|
Artificial Immune Network ; Classification ; ECG ; Pattern Recognition
|
|
Abstract:
|
Electrocardiograph (ECG) is a widely used tool for cardiac monitoring. As traditional ECG curves only represent a short time sampling of patient data, irregular and intermittent cardiac conditions may not be identifiable and therefore real time monitoring is necessary. However, tremendous ECG data are generated by the real time monitoring and immediate health care are needed for the patients, which pose a substantial computational challenge for ECG pattern recognition. Artificial Immune Network (AIN) classifier is proposed to automatically discern and reduce the ECG data size using MemeoryCell representation. This approach combines PCA and K Nearest Neighbor to conduct projection pursuit and supervised classification based on the Memory Cells. Following a real patient study, a simulation study is conducted to exhibit the advantages of our approach compared to traditional methods.
|
