基于改进人工蜂群优化支持向量机的睡眠分期-《中国医学物理学杂志》

文章信息/Info

Title:: Sleep staging using support vector machine optimized by improved artificial bee colony

作者:: 熊馨1; 吴迪1; 张亚茹1; 冯建楠1; 易三莉1; 王春武2; 刘瑞湘3; 贺建峰1; 1.昆明理工大学信息工程与自动化学院，云南昆明 650500； 2.韩山师范学院物理与电子工程学院，广东潮州 521000； 3.云南省第二人民医院临床心理科，云南昆明 650021

Author(s):: XIONG Xin1; WU Di1; ZHANG Yaru1; FENG Jiannan1; YI Sanli1; WANG Chunwu2; LIU Ruixiang3; HE Jianfeng1; 1. College of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, China 2. School of Physics and Electronic Engineering, Hanshan Normal University, Chaozhou 521000, China 3. Department of Clinical Psychology, Yunnan Second Peoples Hospital, Kunming 650021, China

Keywords:: Keywords: sleep staging improved artificial bee colony algorithm support vector machine ReliefF Lévy flight

摘要:: 本研究提出通过改进人工蜂群算法优化支持向量机（IMABC-SVM）进行睡眠分期。对提取的离散小波变换分解数据分量、时域特征、非线性特征、微状态特征，使用ReliefF算法进行特征筛选，提取出最优特征矩阵，并由IMABC-SVM分类器对特征矩阵进行训练。为验证特征筛选与优化分类器效果，进行相关消融实验。结果表明IMABC-SVM方法精度可达89.97%。IMABC-SVM方法可为睡眠相关疾病的检测、预防和治疗提供有效的依据。

Abstract:: A support vector machine (IMABC-SVM) optimized by an improved artificial bee colony algorithm is used to for sleep staging. The data components obtained by discrete wavelet transform, time-domain features, non-linear features and micro-state features are filtered using ReliefF algorithm for obtaining the optimal feature matrix which is then trained by IMABC-SVM classifier. Some ablation experiments are conducted to verify the effectiveness of the feature selection and the optimized classifier. The experimental results show that the accuracy of IMABC-SVM reaches 89.97%. IMABC-SVM can provide a basis for the detection, prevention and treatment of sleep-related disorders.

备注/Memo:: 【收稿日期】2022-11-12 【基金项目】国家自然科学基金（82060329）；云南省科技厅面上项目（202201AT070108）【作者简介】熊馨，博士，讲师，研究方向：主动式脑机接口及模式识别、图像分析与智能辅助，E-mail: xiongxin840826@163.com 【通信作者】贺建峰，博士，教授，研究方向：医学成像仿真与图像处理分析、医疗信息融合与数据挖掘、医疗信息化管理、人工智能和计算机技术应用，E-mail: jfenghe@foxmail.com