The Detection Of VF And AF Signal Based On Nonlinear

In recent years， as the number of old Chinese growing， the number of peoplesuffering from cardiovascular disease increased year by year.This phenomenon hasbecome a great threat for the lives and safety of people， many people will take a heartattack and even he will be death.Malignant arrhythmias is the main cause of suddencardiac death， and this paper is to study ventricular fibrillation (VentricularFibrillation，VF) and atrial fibrillation (Atrial Fibrillation，AF) which are the mostcommon and serious malignant arrhythmias. According to the American HeartAssociation statistics， about30%of the patient has known or unknown heartdisease.Most of the cardiacs do not know they have heart disease incidence，when theyhave a sudden illness，there haven not medical equipment， in this time， we needtake the patient to the hospital，but that requires a lot of time testing to treatment andleading the way to the hospital， they spend so much time that many patients will losetheir lives. If we can detect these malignant arrhythmia earlier， we can greatly improvethe chances of survival of these people. Therefore， the detection of malignantarrhythmia is an important research topic.Ventricular fibrillation (Ventricular Fibrillation， VF) is a kind of seriousarrhythmia.VF can cause rapid heart failure patients to death. VF can not be timelydefibrillation， then the survival chance of patients is greatly reduced， and the patientwill be death in a few minutes.Therefore， the correct and timely detection is veryimportant.Atrial fibrillation (Atrial fibrillation， AF)refers to the occurrence of atrial musclefiber disorderly fibrillation，appear350-600times per minute of uncoordinated，it is anirregular signal.According to statistics，75percent of AF patients complicated withcerebrovascular accidently， it will cause patient disability and even be death suddenly，AF is a serious threat to human health. Therefore， the prediction and diagnose is ofatrial fibrillation， which can greatly improve the quality of treatment， and reduce the incidence and mortality of patients with critical illness， have very importantsignificance to the clinical and social.Support vector machine (Support Vector Machine， SVM) is a machine learningmethod which is based on structural risk minimum principle of statistical theory andthe theory of VC dimension， SVM is a learning machine which is according to thelimited sample of information in the learning ability and complexity seek the bestcompromise， to obtain the best generalization ability. SVM Can identify patternsand analy data， commonly used in the two classification problem. This topicuses support vector machine to achieve nonlinear classification. Because RBF(RBF Radial Basis Function)has many advantages such as it has less modelparameters and wide application， I use RBF for the kernel function of SVM.Simulate signal by using the support vector machine software package of LuZhenbo (SVM_luzhenbo&LS_SVMlab) on matlab software.ECG is nonlinear. In this paper，then we can analy the detection algorithm ofventricular fibrillation and atrial fibrillation by nonlinear method. I use SVM toimplement the algorithm.The steps of this algorithm are as follows：firstly，to get thesignal， secondly， the pretreatment of ECG data obtained denoising and otheroperations， thirdly，to get the number of its feature extraction， and finally make thefeature to support vector machine classifier.The method in my paper can be used in the monitor or mobile phonesoftwarewhich has monitoring function.