Fractal analysis of scaterogram

O. Velychko, Alhalalmeh Sadam Iyad Hamed, I. Mykhailova, T. Kolesnikova
Annotations languages:


Description: The subjects of investigation are long time series and short time series of RR-intervals and their special distribution named as scatterogram. The objective of work is the investigation of one-minute intervals of RR data series to definite the dataset of parameters for estimation of physiological state in real time mode. The methodology of investigation is based on the data proc-essing of the standard indicators of heart rate variability (HRV), fractal analysis, machine learning methods (k-means) and fac-tor analysis. Testing signals were been taken from the European data base of medical signals PhysioNet. Scatterogram analysis of the testing diurnal record with normal sinus rhythm and arrhythmia was performed. A strong correlation between the coeffi-cient of fractal dimension and standard indications of HRV for normal sinus rhythm and null correlation for arythmia had been observed. Standard indicators of HRV and coefficient of fractal dimension for each one-minute interval were calculated and estimated with the Student’s average test that had shown inexpediency of scaterogram surface using. Cluster analysis based on the k-means method to determine general probabilistic data series groups was performed. Initial data for cluster analysis is a matrix of standard HRV indicators, coefficient of fractal dimension and geometric parameters of scatetogram. Entire objects from the one-minute dataset were grouped into six clusters. The average value of scatterogram parameters (coefficient of fractal dimension, long and short axes and their ration) for selected clusters were found with probability more than 95 %. Data set of parameters to analyze the one-minute intervals of RR-dataset was defined with fractal analysis. Proposed dataset of indicators for estimation one-minute intervals includes: pNN50 (the proportion derived by dividing the number of interval differences of successive NN intervals greater than 50 ms by the total number of NN intervals), long SD2 and short SD1 ellipse axes, ratio SD2/SD1, triangular index TI, mean heart rate and coefficient of fractal dimension Df. Gotten results may be used to develop mobile software for personal monitoring devices.


Keywords: heart rate variability, fractal analysis, scatterogram, short time intervals of RR-dataset.

References

Reference:
 Velychko, O.M., Al-Khalalmekh Sadam Eiad Khamed, Mikhailova, E.A. and Kolesnikova, T.A. (2019), “Fraktalnyi analiz skaterohramy” [Fractal analysis of scaterogram], Information Processing Systems, Vol. 3(158), pp. 42-53. https://doi.org/10.30748/soi.2019.158.05.

Whoops, looks like something went wrong.