Evolutionary algorithms for multiobjective and multimodal optimization of diagnostic schemes |
| |
Authors: | de Toro Francisco Ros Eduardo Mota Sonia Ortega Julio |
| |
Affiliation: | Department of Signal Theory, Telematics and Communications, ETS Informática, Granada, Spain. ftoro@ugr.es |
| |
Abstract: | This paper addresses the optimization of noninvasive diagnostic schemes using evolutionary algorithms in medical applications based on the interpretation of biosignals. A general diagnostic methodology using a set of definable characteristics extracted from the biosignal source followed by the specific diagnostic scheme is presented. In this framework, multiobjective evolutionary algorithms are used to meet not only classification accuracy but also other objectives of medical interest, which can be conflicting. Furthermore, the use of both multimodal and multiobjective evolutionary optimization algorithms provides the medical specialist with different alternatives for configuring the diagnostic scheme. Some application examples of this methodology are described in the diagnosis of a specific cardiac disorder-paroxysmal atrial fibrillation. |
| |
Keywords: | |
本文献已被 PubMed 等数据库收录! |
|