吉林大学学报(理学版) ›› 2021, Vol. 59 ›› Issue (1): 101-106.

• • 上一篇    下一篇

基于机器学习的跨患者癫痫自动检测算法

杨舒涵1,2, 李博2,3, 周丰丰1,2   

  1. 1. 吉林大学 计算机科学与技术学院, 长春 130012;2. 吉林大学 符号计算与知识工程教育部重点实验室, 长春 130012;
    3. 吉林大学 软件学院, 长春 130012
  • 收稿日期:2020-02-05 出版日期:2021-01-26 发布日期:2021-01-26
  • 通讯作者: 周丰丰 E-mail:ffzhou@jlu.edu.cn

Automatic Epileptic Seizure Detection Algorithm for Non-specific Patient Based on Machine Learning

YANG Shuhan1,2, LI Bo2,3, ZHOU Fengfeng1,2   

  1. 1. College of Computer Science and Technology, Jilin University, Changchun 130012, China;
    2. Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, China;
    3. College of Software, Jilin University, Changchun 130012, China
  • Received:2020-02-05 Online:2021-01-26 Published:2021-01-26

摘要: 针对目前癫痫自动检测算法多集中于为单个患者建立检测模型, 泛化能力较弱的问题, 提出一种基于机器学习的跨患者癫痫自动检测算法. 该算法使用多个癫痫患者的脑电数据, 先对数据进行预处理后分析脑电数据间存在的特征, 再对特征进行筛选, 训练出一个跨患者的癫痫自动检测模型. 该算法不需为每个患者
建立单独的检测模型, 实现了仅使用一个检测模型即可对不同患者进行癫痫检测. 实验结果准确率为0.877 4, 敏感性为0.854 8, 特异性为0.9.

关键词: 癫痫检测, 机器学习, 脑电数据, 滤波器, 特征提取, 特征选择

Abstract: Aiming at the problem that the automatic epileptic seizure detection algorithm focused on the establishment of a detection model for single patient, and the generalization ability was weak, we proposed an automatic epileptic seizure detection algorithm for non-specific patient based on machine learning. The algorithm used the electroencephalography (EEG) data of multiple epileptic patients, analyzed the characteristics of EEG data after preprocessing the data, and then selected the characteristics to train an automatic epileptic seizure detection model for non-specific patients. The algorithm did not need to establish a separate detection model for each patient, it could detect epilepsy in different patients with only one detection model. The accuracy, sensitivity and specificity of the algorithm are 0.877 4,0.854 8 and 0.9, respectively.

Key words: seizure detection, machine learning, electroencephalography data, filter, feature extraction, feature selection

中图分类号: 

  • TP399