Tachycardia Discrimination in Implantable Cardioverter Defibrillators Using Support Vector Machines and Bootstrap Resampling
Support Vector Machines: Theory and ApplicationsStudies in Fuzziness and Soft Computing(2005)
摘要
Accurate automatic discrimination between supraventricular (SV) and ventricular (V) tachycardia (T) in implantable cardioverter defibrillators (ICD) is still today a challenging problem. An interesting approach to this issue can be Support Vector Machine (SVM) classifiers, but their application in this scenario can exhibit limitations of technical (reduced data sets are only available) and clinical (the solution is inside a hard-to-interpret black box) nature. We first show that the use of bootstrap resampling can be helpful for training SVM with few available observations. Then, we perform a principal component analysis of the support vectors that leads to simple discrimination rules that are as effective as the black box SVM. Therefore, a low computational burden method can be stated for discriminating between SVT and VT in ICD.
更多查看译文
关键词
principal component analysis,bootstrap resampling,support vector,defibrillator,tachycardia
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络