Design anIntelligent Ballistocardiogr aphic Chairusing Novel QuickLearn andSF-ARTAlgorithms andBiorthogonal Wavelets

Todesign aheartdiseases diagnosing system, we applied compactly supported Biorthogonal wavelet transform toextract essential features oftheBallistocardiogram (BCG) signal andtoclassify themusing twonovelsupervised learning algorithms called SF-ARTandQuickLearn. Initial tests with BCG fromsixsubjects (both healthy andunhealthy people) indicate thatbothSF-ARTandQuicklearn algorithms can classify thesubjects intothreeclasses withhighaccuracies, highlearning speeds, and verylowcomputational loads comparedto thewell-known neuralnetworkssuchas Multilayer Perceptrons. The proposedheartdiseases diagnosing systems arealmost insensitive tolatency andnon- linear disturbance. Moreover, thewavelet transform requires no priorknowledge ofthestatistical distribution ofdata samples andthecomputational complexity andtraining time arereduced. I.INTRODUCTION B allistocardiogra 0 ¢~ ~ ~~~T - -- -