基于模糊支持向量机的高速公路交通事件的自动检测  

Approach to Automatic of Highway Detection Incident Based on Fuzzy support vector machine

作　　者：焦军彩[1] 

机构地区：[1]南京工业大学数学系,江苏南京210009

出　　处：《盐城工学院学报（自然科学版）》2009年第4期70-72,共3页Journal of Yancheng Institute of Technology:Natural Science Edition

摘　　要：利用支持向量机的全局优化、适应性强、泛化性能好等优点,针对实时交通流数据的随机性、高维、非线性和时变等特性,将模糊支持向量机应用于高速公路交通事件检测问题中。在识别阶段利用60组实测数据训练模糊支持向量机,利用60组实测数据进行测试,测试结果表明,利用FSVM进行交通事件检测,识别率达到96.7%。Support Vector Machine（SVM） has the advantages of global solutions, good adaptability, high generalization in theory. Due to the randomicity , high dimension, nonlinear, time -variant of the traffic flow data, Fuzzy support vector machine （FSVM） applied to high way incident detection In the stage of traffic incident recognition, sixty train data was used to train the FSVM, sixty test data was used to testify the effect of the recognition model. The results show that the recognition rate is approximate to 96.7%, which also proved the feasibility of the approach proposed in this paper.

关 键 词：模糊支持向量机 交通事件自动检测 隶属函数 

分 类 号：U495[交通运输工程—交通运输规划与管理]