以長時間的心律變異數量化分析不規律心跳的發生頻率

Quantifying the Occurring Frequency of Irregular Heartbeats with Long-term Heart Rate Variability

洪睿謙(Jui-Chien Hung)

指導教授 ： 劉省宏

朝陽科技大學/資訊學院/資訊工程系/碩士(2014年)

摘要

心電圖信號包含了可以幫助醫生診斷的重要訊息，如果是正常或衰竭的心臟，心電圖都會有一定的特徵存在，不正常的心跳，稱之為心律不整，一般人都會偶然發生幾次的心律不整的心跳，這對人並不會有太大的影響，但若是心律不整的發生頻率增高，這是一種警訊，患者發生休克或猝死的風險度就會增高。本研究目的是探討利用心律變異數量化分析不規律的心律不整發生頻率，設定25%和10%的發生頻率，當超過設定時，視為心臟處於高風險。本研究使用美國麻省理工學院所提供之心律不整資料庫做為資料來源，為了評估心律變異數特徵的鑑別能力，我們採用2個分類器分別為支持向量機與決策樹，利用心律變異數的十種時域與頻域參數為風險評估的特徵，利用支持向量機和決策樹分類器透過十倍交叉驗證法，來得到最佳心律變異數的參數。在10%和25%的發生頻率，結果顯示支持向量機分類器在評估心臟功能的風險程度上有較佳的表現，準確度為92.0% 和95.3%。由本研究可知，以心律變異數分析不規律的心律不整的心跳，可以評估出心臟功能的風險程度，此方法有益於對那些已知其有心律不整的患者，能評估其心臟功能風險度是否有增加。

關鍵字

支持向量機 ； 心律不整 ； 心電圖 ； 決策樹

並列摘要

Electrocardiogram (ECG) signal contains important information that can help doctors diagnose. If it is normal or failure of the heart, ECG has some characteristics, irregular heartbeats, called arrhythmias. Generally arrhythmic people are by accident a few arrhythmic heartbeats who do not have much of an impact. However, if the frequency of arrhythmias increases, it is a warning, patients with increased risk of shock or sudden death will. The aim of this study was to investigate the use of heart rate variability (HRV) for the quantitative analysis of irregular arrhythmias incidence, set the frequency of 25% and 10%, when more than setting, as a heart is at high risk, as abnormal condition. This study used United States Massachusetts Institute of technology offers an arrhythmia database as a data source. In order to assess the identifying performance of the characteristics of HRV, we used 2 decision trees and support vector machine (SVM) classifier respectively which input features were time and frequency domain parameters of heart rate variability to identify the normal or abnormal condition. The classifiers by 10-fold cross-validation got the optimum parameters of HRV. The results show that support vector machine has the best accuracy of 95.3% when irregular arrhythmia rate is set at 25%. However, when irregular arrhythmia rate is set at 10%, the results show that support vector machine has the best accuracy was 92.0%.The present study showed that, we can evaluate the degree of risk of cardiac function by the HRV. This method is beneficial for those who know their patients with arrhythmias, to assess their heart function risk will increase.

並列關鍵字

Support vector machines ； Arrhythmia ； Electrocardiogram ； Decision trees

參考文獻

[2] 彰化基督教醫院， http://www2.cch.org.tw/intern/ugy%20lecture/

[7] M. Bashir , D. Lee , M. Li , H. Shon , J. Bae , M. Cho and K. Ryu, "Trigger learning and ECG parameter customization for remote cardiac clinical care information system," IEEE Trans. Inf. Technol. Biomed, to be published, DOI: 10.1109/TITB.2012.2188812 .

[8] M.G. Tsipouras, D.I. Fotiadis, and D. Sideris, "An arrhythmia classification system based on the RR-interval signal," Artificial Intelligence in Medicine, vol. 33, Issue 3, pp. 237-250, 2005.

[9] Bernardi et al., L. Bernardi, J. Wdowczyk-Szulc, C. Valenti, et al, "Effects of controlled breathing, mental activity and mental stress with or without verbalization on heart rate variability," J. Am. Coll. Cardiol, pp. 1462-1469, 2000.

[10] B.M. Asl, S.K. Setarehdan and M. Mohebbi, "Support vector machine-based arrhythmia classification using reduced features of heart rate variability signal," Artificial intelligence in medicine, vol. 44, Issue 1, pp. 51-64. 2008.