%0 Journal Article
%T 基于权值优化的FCM-MSVM算法及其在高速公路状态判别中的应用<br>FCM-MSVM algorithm based on weight optimization and its application in state identification of freeway
%A 李晓璐
%A 于昕明
%A 杜崇
%A 张溪
%A 张彭
%A 朱广宇
%J 北京交通大学学报
%D 2018
%R 10.11860/j.issn.1673-0291.2018.04.010
%X 摘要 高速公路实时交通状态的准确判别是科学制定交通管理策略的重要基础.本文以实测的高速公路交通流三特征参数的数据作为输入，引入熵权法刻画参数之间重要程度的差异，利用改进的FCM算法对特征参数进行聚类，根据不同交通状态的结果，建立基于多分类器支持向量机的交通状态判别模型，并分别采用网格搜索法、遗传算法和粒子群算法对支持向量机参数进行优化，从而提高模型的判别准确率.最后选取实际数据对本文模型进行验证，判别结果的分类准确率可达96.3980%.<br>Abstract：The accurate identification of expressway real-time traffic status is an important basis for scientifically formulating traffic management strategy. In this paper, the measured data of three traffic parameters of freeway traffic flow are taken as input. The entropy weight method is introduced to characterize the difference of importance between the parameters. The improved FCM algorithm is used to cluster the characteristic parameters. Based on the results of different traffic states, Multi-classification SVM traffic state discriminant model is constructed, and using grid search method, genetic algorithm and particle swarm optimization algorithm to optimize the parameters of support vector machine to improve the accuracy of the model. Finally, the actual data is selected to verify the model. The accuracy of the classification results can reach 96.3980%.
%K 交通状态判别
%K 高速公路
%K 模糊均值聚类
%K 多分类支持向量机
%K 参数寻优&lt
%K br&gt
%K traffic state identification
%K expressway
%K fuzzy mean clustering
%K multi classification support vector machine
%K parameter optimization
%U http://jdxb.bjtu.edu.cn/CN/abstract/abstract3401.shtml