Extracting Features for Cardiovascular Disease Classification Based on Ballistocardiography

Cardiovascular disease affects the health of people seriously in the world, especially in the elderly. This paper proposes an effective approach of detecting and analyzing the health status of the elderly. In this work, we continuously acquire Ballisto cardiography (BCG) signal with the micro-movement sensitive mattress (MSM) during non-intrusive sleep in home environment. In the paper, we propose a new method to extract heartbeat intervals (RR) based on Ensemble Empirical Mode Decomposition (EEMD), and extract the signal features by calculating the parameters of heart rate variability (HRV) from time domain analysis, frequency domain analysis and nonlinear analysis. A Naïve Bayesian Classification method is applied to classify the normal persons, hypertension patients and coronary heart disease (CHD) patients by using the obtained features. The proposed method is evaluated by using the BCG datasets from eighteen subjects, including eight females and ten males (age 40-72). The results are satisfactory and can provide a classification precision of 92.3%.