Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (23): 131-135.DOI: 10.3778/j.issn.1002-8331.1808-0333

Previous Articles     Next Articles

Research on Collaborative Optimization of Gait Characteristics and SVM Parameters Based on GA

TANG Chen, XU Shengqiang, CHENG Nan, DUAN Zhikui, WU Xi   

  1. 1.Institute of Industrial and Equipment Technology, Hefei University of Technology, Hefei 230009, China
    2.Institute of Intelligent Machines, Chinese Academy of Sciences, Hefei 230031, China
    3.Hospital Affiliated to Institute of Neurology, Anhui University of Chinese Medicine, Hefei 230061, China
    4.Foshan University, Foshan, Guangdong 528000, China
    5.School of Computer and Information, Hefei University of Technology, Anhui 230009, China
  • Online:2019-12-01 Published:2019-12-11

基于GA的步态特征与SVM参数协同优化研究

唐晨,许胜强,程楠,段志奎,吴玺   

  1. 1.合肥工业大学 工业与装备技术研究院,合肥 230009
    2.中国科学院 合肥智能机械研究所,合肥 230031
    3.安徽中医药大学 神经病学研究所附属医院,合肥 230061
    4.佛山科学技术学院,广东 佛山 528000
    5.合肥工业大学 计算机与信息学院,合肥 230009

Abstract: To assist doctors in better diagnosis of patients with Parkinson’s disease, genetic algorithm is used to optimize the gait characteristics and support vector machine parameters. In the feature selection section, the gait features and support vector machine parameters are generated in a binary coded manner and input into the genetic algorithm. Taking into account the search performance and runtime of the genetic algorithm, the two parameters of the population size and number of evolutionary generation in the algorithm are optimized to find the appropriate population size and number of evolutionary generation. The experimental results show that the average accuracy rate is 85.11%, compared with other feature selection algorithms, the accuracy rate is up to 14% higher. It shows that the use of this method can effectively remove the gait characteristics of patients with Parkinson’s disease, which is beneficial to the diagnosis of patients.

Key words: genetic algorithm, feature selection, Parkinson’s disease, population size, number of evolutionary generation

摘要: 为了辅助医生对帕金森病患者进行更好的诊断,采用遗传算法对患者步态特征和支持向量机参数进行协同优化。在特征选择部分将步态特征和支持向量机参数使用二进制编码的方式生成染色体,同时输入到遗传算法中。考虑到遗传算法的搜索性能和运行时间,针对算法中的种群大小和进化代数这两个参数进行了寻优,找出合适的种群大小和进化代数。实验结果显示平均准确率为85.11%,相对于其他特征选择算法,准确率最高高出14%。表明使用该方法后对帕金森病患者的步态特征进行了有效去除,有利于患者的诊断。

关键词: 遗传算法, 特征选择, 帕金森病, 种群大小, 进化代数