Authors:
Vladimir Kublanov
1
;
Anton Dolganov
1
and
Yan Kazakov
2
Affiliations:
1
Ural Federal University, Russian Federation
;
2
Ural State Medical University, Russian Federation
Keyword(s):
Heart Rate Variability, Arterial Hypertension, Classification, Discriminant Analysis.
Related
Ontology
Subjects/Areas/Topics:
Applications and Services
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Cardiovascular Signals
;
Computer Vision, Visualization and Computer Graphics
;
Medical Image Detection, Acquisition, Analysis and Processing
Abstract:
The paper presents investigation of the diagnostic possibilities of the arterial hypertension using linear and quadratic combinations of the heart rate variability signals features. For this study, two groups were considered: healthy volunteers and patients suffering from the arterial hypertension of the II-III degree. For the study, features of statistical, geometric, spectral, nonlinear and multifractal methods were investigated. Results of the analysis have shown that among studied combinations four feature sets (heart rate, features of the VLF frequency band and LF/HF ratio) have the highest classification accuracy – 93%.