摘要
针对轨道交通司机驾驶水平定量化评价问题,笔者在数据提取、数据降噪和数据降维的基础上,构建了基于支持向量机的轨道司机驾驶水平评价模型方法对该问题进行了研究。研究结果表明:选用高斯核函数的SVM模型在准确率和稳定性上要优于普通、线性和多项式SVM模型,和人工评价结果比较其余弦相似度均高于0.98,模型评价结果的有效性。
Aiming at the quantitative evaluation of the driving level of rail transit drivers,based on data extraction,data denoising and data dimensionality reduction,an evaluation model of rail drivers' driving level based on support vector machine was established and studied.The research results show that the accuracy and stability of SVM model using Gaussian kernel function is better than that of ordinary,linear and polynomial SVM models.Compared with the manual evaluation results,the cosine similarity of SVM model with Gaussian kernel function is higher than 0.98,which indicates the validity of evaluation results of the proposed model.
作者
刘杰
LIU Jie(Chongqing Vocational Institute of Engineering,School of Intelligent Manufacturing and Transportation,Chongqing 402260,China)
出处
《重庆交通大学学报（自然科学版）》
CAS
CSCD
北大核心
2020年第10期31-36,共6页
Journal of Chongqing Jiaotong University(Natural Science)
基金
国家自然科学基金项目(61703351)
五邑大学校内科研项目(2018AL033)。
关键词
交通工程
司机驾驶水平评价
行车数据
小波降噪
主成分分析
支持向量机
余弦相似度
traffic engineering
driver driving level evaluation
driving data
wavelet denoising
principal component analysis
support vector machine
cosine similarity
