Minimum Distance Classifiers in Coronary Artery Disease Diagnosing

• Autorzy: Lewenstein K. , Chojnacki M.
• Konferencja: International Conference on Modelling in Mechatronics (9 ; 02-04.09.2004 ; Kazimierz Dolny, Poland)
• Języki publikacji: EN

Abstrakty

EN

Results of application "minimum distance classifiers" buillt according the "k-nearest neighours" and "set's space covering", created without and with using several "feature selection" strategies for diagnosis of coronary artery disease on base of ECG exercise test have been presented in the paper. Correctness of the patient state diagnosis for the best classifier presented below estimated by "leave one out" method came to 99.8%.

Słowa kluczowe

EN

minimum distance classifiers; feature selection; coronary artery disease; exercise test ECG; ECG; k-nearest neighbours

• Wydawca: Oficyna Wydawnicza Politechniki Warszawskiej
• Czasopismo: Machine Dynamics Problems
• Rocznik: 2004
• Tom: Vol. 28, No. 4
• Strony: 27--33
• Opis fizyczny: Bibliogr. 5 poz.

Twórcy

autor

Lewenstein K.

• Warsaw University of Technology, Institute of Precision and Biomedical Engineerig

autor

Chojnacki M.

• Warsaw University of Technology, Institute of Precision and Biomedical Engineerig

Bibliografia

• Froelicher, V., 1999, Exercise Tests Manual, Warsaw, Bell Corp.
• Janecki, J., 1996, Optimisation of Features Selection Methods in Multi-Dimensional Diagnostic Systems, IBIB Works, Warsaw.
• Jóźwik, A., 1994, Pattern Recognition Method Based on k-Nearst Neighbours Rules, Journal on Communications, 45.
• Jóźwik, A., Janecki, J., Demczuk, M., 1998, A New Type of Two Stage NN Clasifier and NN Clasifiers Nets. Application to Cardio-Cilculatory Events Prediction, Proc. of VIII Int. Conf. IMEKO, Dubrovnik, 11.9-11.12.
• Lewenstein, K., 2002, Artificial Neural Networks in the Diagnostics of Coronary Artery Disease Based on ECG Exercise Test, Oficyna Wydawnicza PW, Warsaw.