基于粒子群优化支持向量机的癫痫发作预测  

Epileptic Seizure Prediction Based on Support Vector MachineOptimized by Particle Swarm Optimization

作　　者：刘金森 黄炜嘉[1] 李效龙[1] LIU Jinsen;HUANG Weijia;LI Xiaolong(College of Electronic Information,Jiangsu University of Science and Technology,Zhenjiang 212100)

机构地区：[1]江苏科技大学电子信息学院,镇江212100

出　　处：《计算机与数字工程》2023年第3期736-741,747,共7页Computer & Digital Engineering

基　　金：国家自然科学基金面上项目(编号:61671221)资助。

摘　　要：为实现癫痫发作自动预测,采用一种基于粒子群(PSO)与支持向量机(SVM)方法相结合的癫痫发作预测模型。首先利用小波包变换提取小波包系数相对能量特征,接着将训练样本输入到支持向量机中,经过粒子群算法寻找最优参数从而实现对测试样本进行分类识别,最后论文提出一种阈值决策方法对支持向量机输出结果进行后处理。对CHB-MIT癫痫数据库中5名癫痫病人进行研究,通过与随机参数法、网格搜索方法对比分析,发现使用粒子群优化支持向量机提高了发作前期识别准确率,减少了误警率,从而取得更好的预测效果,证明了该方法的可行性,对癫痫预测研究具有一定的应用价值。In order to realize the automatic prediction of epileptic seizures,an epilepsy prediction model based on particle swarm optimization and support vector machine is proposed.Firstly,wavelet packet transform is used to extract the relative energy characteristics of wavelet packet coefficients.Then,the training samples are input into the support vector machine,and the optimal parameters are found by particle swarm optimization.Finally,a threshold decision is proposed to process the output of support vec⁃tor machine.Five epileptic patients in CHB-MIT epilepsy database are studied.Compared with random parameter method and grid search method,it is found that PSO-SVM improved the accuracy of preictal recognition,the false alarm rate is reduced,and bet⁃ter prediction effect is obtained.It proves the feasibility of the method and has certain application value for epilepsy prediction re⁃search.

关 键 词：癫痫预测 小波包变换 支持向量机 粒子群算法 

分 类 号：R318[医药卫生—生物医学工程] TP391[医药卫生—基础医学]