Multi-objective Optimization for Paroxysmal Atrial Fibrillation Diagnosis

This paper deals with the application of multi-objective optimization to the diagnosis of Paroxysmal Atrial Fibrillation (PAF). The automatic diagnosis of patients that suffer PAF is done by analysing Electrocardiogram (ECG) traces with no explicit fibrillation episode. This task presents difficult problems to solve, and, although it has been addressed by several authors, none of them has obtained definitive results. A recent international initiative to study the viability of such an automatic diagnosis application has concluded that it can be achieved, with a reasonable efficiency. Furthermore, such an application is clinically important because it is based on a non-invasive examination and can be used to decide whether more specific and complex diagnosis testing is required. In this paper we have formulated the problem in order to be approached by a multi-objective optimisation algorithm, providing good results through this alternative.