Vector of Diagnostic Features in the Form of Decomposition Coefficients of Statistical Estimates Using a Cyclic Random Process Model of Cardiosignal

Abstract

The paper suggests the use of diagnostic features in the form of decomposition coefficients of electrocardiograms statistical estimates in the normal range and in different types of pathologies obtained on the basis of a mathematical model in the form of a cyclic random process. As a criterion for choosing the necessary spectral coefficients of the cardiosignal decompositions in the Chebyshev basis set, was chosen the energy criterion. As diagnostic features, was used a such spectral coefficients that according to the Bessel inequality, contribute to the energy of the cardiosignal statistical estimation realization not less than 95% at their minimal number. Diagnostic spaces were modeled to present the delimitation and grouping of diagnostic features and the distances between the centers of spectral coefficient groups for the normal range and in pathology were found.