Mining Biosignal Data: Coronary Artery Disease Diagnosis Using Linear and Nonlinear Features of HRV
The main purpose of our study is to propose a novel methodology to develop the multi-parametric feature including linear and nonlinear features of HRV (Heart Rate Variability) diagnosing cardiovascular disease. To develop the multi-parametric feature of H
- PDF / 1,003,478 Bytes
- 11 Pages / 430 x 660 pts Page_size
- 42 Downloads / 189 Views
Database/Bioinformatics Laboratory, School of Electrical & Computer Engineering, Chungbuk National University, Cheongju, 361-763, Korea {hglee,khryu}@dblab.chungbuk.ac.kr 2 Korea Research Institutes of Standards and Science, Korea [email protected]
Abstract. The main purpose of our study is to propose a novel methodology to develop the multi-parametric feature including linear and nonlinear features of HRV (Heart Rate Variability) diagnosing cardiovascular disease. To develop the multi-parametric feature of HRV, we used the statistical and classification techniques. This study analyzes the linear and the non-linear properties of HRV for three recumbent positions, namely the supine, left lateral and right lateral position. Interaction effect between recumbent positions and groups (normal and patients) was observed based on the HRV indices and the extracted HRV indices used to classify the CAD (Coronary Artery Disease) group from the normal people. We have carried out various experiments on linear and nonlinear features of HRV indices to evaluate several classifiers, e.g., Bayesian classifiers, CMAR, C4.5 and SVM. In our experiments, SVM outperformed the other classifiers.
1 Introduction Mortality of domestic people from cardiovascular disease ranked second, which followed that of from cancer last year. Therefore, it is very important and urgent to enhance the reliability of medical examination and treatment for cardiovascular disease. HRV (Heart Rate Variability) analysis has been used extensively to assess autonomic control of the heart under various physiological and pathological conditions, and used as a clinical tool to diagnose cardiac autonomic function [1], [2].Various measures and explanations have been used to analyze the HRV. For example, simple linear time domain analysis, such as mean, standard deviation, and root mean square of successive RR interval differences have been widely employed in quantification of the overall variability of the Heart Rate (HR) [2]. Frequency domain variable provide *
**
This work was supported by the Korea Research Foundation Grant funded by the Korean Government (The Regional Research Universities Program/Chungbuk BIT ResearchOriented University Consortium). Corresponding author.
T. Washio et al. (Eds.): PAKDD 2007 Workshops, LNAI 4819, pp. 218–228, 2007. © Springer-Verlag Berlin Heidelberg 2007
Mining Biosignal Data: Coronary Artery Disease Diagnosis
219
markers of the cardiac autonomic regulation, i.e. the sympatho-vagal balance [3], [4]. In addition, there are several nonlinear features. The nonlinear interaction between the various regulatory systems of the heart rate gives rise to clinically useful concepts of variability and regularity. The complexity of the human physiological system, which is reduced in bad health but increased in good health [5], can be analyzed quantitatively by various nonlinear methods. Nonlinear analyses include the Princare plots and complexity estimation. Therefore, we consider it worth-while investigating the linear and nonlinear prop
Data Loading...
