Knowledge-based and Data-driven Models in Arrhythmia Fuzzy Classification

Summary

Objectives: Fuzzy rules automatically derived from a set of training examples quite often produce better classification results than fuzzy rules translated from medical knowledge. This study aims to investigate the difference in domain representation between a knowledge-based and a data-driven fuzzy system applied to an electrocardiography classification problem.

Methods: For a three-class electrocardiographic arrhythmia classification task a set of fifteen fuzzy rules is derived from medical expertise on the basis of twelve electrocardiographic measures. A second set of fuzzy rules is automatically constructed on thirty-nine MIT-BIH database’s records. The performances of the two classifiers on thirteen different records are comparable and up to a certain extent complementary. The two fuzzy models are then analyzed, by using the concept of information gain to estimate the impact of each ECG measure on each fuzzy decision process.

Results: Both systems rely on the beat prematurity degree and the QRS complex width and neglect the P wave existence and the ST segment features. The PR interval is not well characterized across the fuzzy medical rules while it plays an important role in the data-driven fuzzy system. The T wave area shows a higher information gain in the knowledge based decision process, and is not very much exploited by the data-driven system.

Conclusions: The main difference between a human designed and a data driven ECG arrhythmia classifier is found about the PR interval and the T wave.

Keywords

Arrhythmia - Fuzzy Logic - Artificial Intelligence

• References

• 1 Cherkasskj V, Mulier F. Learning from data. John Wiley and Sons Inc; 1998
• 2 Liu H, Motoda H. eds. Feature Extraction, Construction and Selection: A Data Mining Perspective. Kluwer International Series in Engineering and Computer Science. 1998
• 3 Zadeh LA. A fuzzy-algorithmic approach to the definition of complex or imprecise concepts. Int J Man-Machine Studies 1976; 8: 249-91.
• 4 Degani R. Computerized electrocardiogram diagnosis: fuzzy approach. Methods Inf Med 1992; 31 (Suppl. 84) 225-33.
• 5 Presedo J, Vila J. et al. Fuzzy modeling of the expert’s knowledge in ECG based ischemia detection. Fuzzy Sets and Systems 1996; 77 (Suppl. 01) 63-75.
• 6 Stefanelli M, Bellazzi R. et al. A general architecture for medical expert systems (GAMES). In: Noortho van van Goor J, Christensen JP. eds. AIM GAMES Advances in medical Informatics. IOS Press; 1992
• 7 MIT-BIH database distributor.. Beth Israel Hospital, Biomedical Engineering, Division KB-26, 330 Brookline Avenue, Bostin, MA 02215, USA.
• 8 Silipo R, Berthold M. Discriminative Power of Input Features in a Fuzzy Model. In: Advances in Intelligent Data Analysis (LNCS 1642). Hand D, Kok J, Berthold M. eds. Springer-Verlag Springer-Verlag; 1999. -87.
• 9 Zong W, Jiang D. Automated ECG rhythm analysis using fuzzy reasoning. Proceedings of Computers in Cardiology. 1998: 69-72.
• 10 Silipo R, Marchesi C. Artificial Neural Networks for automatic ECG analysis. IEEE Transactions on Signal Processing 1998; 46 (Suppl. 05) 1417-25.
• 11 Silipo R. Extracting Information from Fuzzy Models. Proceedings of NAFIP, to appear,. 2000
• 12 De Chazal P, Celler B. Selection of Optimal Parameters for ECG Diagnostic Classification. Proceedings of Computers in Cardiology. 1997: 13-6.
• 13 Berthold MR, Huber KP. Building Precise Classifiers with Automatic Rule Extraction. Proceedings of IEEE International Conference on Neural Networks 1995; 3: 1263-8.