摘要
癫痫检测一直是医学界的难题,对于检测算法而言,分类精度和敏感度是衡量检测系统的重要指标。癫痫检测实际上可归分为二分类问题,对开源数据集CHB-MIT Scalpe EEG Database的23例患者单导联C3-P3,利用BP、CNN和SVM网络进行癫痫检测探究。检测结果表明,BP网络、CNN网络、SVM分类器的分类进度分别为67.01%、83.33%、93.40%,在未来癫痫检测分类应用方面具有较大的潜在意义。
Epilepsy detection has always been a difficult problem in the medical community.For the detection algorithm,accuracy and error rate are important indexes to measure the detection system.Epilepsy detection can be classified into two classification problem actually,in this paper,the open source data collection of CHB-MIT Scalpe EEG Database of 23 cases with single lead C3-P3,using BP,CNN and the SVM network for automatic detection.The test results show that the BP network,the network,the accuracy of the SVM classifier are 63.8%,80.2%and 93.78%respectively.It has great potential significance in the future application of classification of epilepsy detection.
出处
《工业控制计算机》
2019年第1期79-80,共2页
Industrial Control Computer
基金
国家自然科学基金资助项目(60673132)
