HySMID: An Ischemia Diagnosis System Using Genetic Algorithms and Multicriteria Decision Analysis

Multicriteria decision analysis is employed in a fourstage (preprocessing, beat classification, window characterization and episode definition) ischemia diagnosis system, named HySMID. Beat classification is realized using a multicriteria approach where each beat is compared to already classified category prototypes and the similarity is computed in a fuzzy way. The comparison is based upon five criteria that include ST segment changes, T wave alterations and patient’s age. The difficulty in applying these criteria is the determination of the required parameter values (thresholds and weights). To overcome it, a genetic algorithm is used, which after proper training, automatically calculates the optimum values of the above parameters. HySMID was validated using the European Society of Cardiology ST-T database and performed better than previously reported methods.