A Data Mining Approach and Framework of Intelligent Diagnosis System for Coronary Artery Disease Prediction

摘要

The death rates from diseases in a recent medical report are much higher than from accidents and disasters. Coronary heart disease has the top rank in causing deaths in the United States. Statistical reports show that 130.7 (cancer), 77.2 (cerebrovascular disease), and 37.2 (cardiovascular disease) per 100,000 people in the Korean population have been dying as the result of the diseases every year. In addition, increased carotid intima-media thickness (IMT) is associated with atherosclerosis risk factors and adverse cardiovascular outcomes. Hence, it is very important to detect early symptoms of heart problems as this detection can result in better treatments with better results. It is also important to develop various diagnostic features that can help in diagnosing cardiovascular disease based on pathological characteristics of Korean people to enhance the reliability of a diagnosis. In this study, we suggest that the various features, takes into consideration the whole possible linearonlinear features of heart rate variability and carotid arterial wall thickness, may be helpful to diagnose the cardiovascular disease as a diagnostic supplementary tool. Heart rate variability (HRV) has been used extensively to assess autonomic control of the heart under various physiological and pathological conditions. In cardiology, HRV is used as a clinical tool to diagnose cardiac autonomic function. Various features have been used to analyze HRV. The linear features of HRV provide markers of cardiac autonomic regulation. 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. Nonlinear analyses include the complexity estimation, the fractal scaling analysis such as the Hurst exponent, and detrended fluctuation analysis, approximate entropy, etc.