KNN方法在高速公路多义性通行费拆分中的应用  

Study on Ambiguity Toll Distribution for Expressway with KNN Theory

作　　者：王秀丽[1] 朱耿先[2] 

机构地区：[1]天津理工大学管理学院,天津300191 [2]天津高速公路集团有限公司,天津300384

出　　处：《北京理工大学学报（社会科学版）》2012年第5期45-49,共5页Journal of Beijing Institute of Technology：Social Sciences Edition

基　　金：国家社会科学基金资助项目(08BJY004);天津市高等学校人文社会科学研究项目(20102103)

摘　　要：正确判断匹配样本状态并采用科学合理的拆分方法是高速公路多义性通行费拆分的关键。利用K近邻聚类(KNN)的基本理论进行匹配样本状态的判别,给出高速公路多义性通行费拆分方法。以天津市京沪高速和津沧高速二义性路径问题为例,比较应用KNN和RBF神经网络及BP神经网络匹配样本状态的判别效果。研究表明:应用KNN的高速公路多义性通行费拆分方法较RBF、BP神经网络法更客观、公平。The key to ambiguity toll distribution for expressway is to judge the matching states and use the scientific distribution method. This paper studied the differentiation of matching sample by using K-Nearest Neighbor （KNN） theory, and put forward the algorithm of ambiguous toll distribution based on KNN and license plate recognition technology. Taking the ambiguity of Beijing- Shanghai expressway and Tianjin-Changzhou expressway as an example,this paper proved that the discriminant effect of KNN is better than RBF neural network and BP neural network by experiments, and then proved that the distribution method for ambiguous toll of expressway based on KNN is fairer and more objective.

关 键 词：K近邻聚类(KNN) 车牌识别 路径多义性 RBF神经网络 BP神经网络 

分 类 号：F224[经济管理—国民经济]