文章基本信息

标题：One-Class Learning Based Algorithm for the Freeway Automatic Incident Detection
本地全文：下载
作者：Zhiyong Liu ; Menghua Zhu ; Keqing Fan 等
期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2006
卷号：6
期号：10
页码：289-293
出版社：International Journal of Computer Science and Network Security
摘要：According to the characteristic of the freeway, choosing three the most sensitive traffic parameters of AID(Automatic Incident Detection) as the feature vectors, an one-class classification based AID algorithm is proposed. The method establishes the regional distribution model of learning samples, and constructs the decision function in the feature space. When the detected data fall into the inner of the decision region, it will be judged as no incident happened, or else incidents happened. This method does not require any transcendental statistical hypothesis about the distribution of samples, even if the distribution of samples is non-convex and unconnected, it can gain better decision function. The experiment results show that the new algorithm can obtain a higher IR, and limit FIR effectively, so it is a new and potential method for AID
关键词：AID, One-class Classification,SVM,Freeway