基于k-mediods及其改进算法的非法营运车辆识别  

RECOGNISING ILLEGAL OPERATION VEHICLES BASED ON K-MEDIODS AND ITS IMPROVED ALGORITHM

作　　者：帅丹[1] 蓝章礼[1] 李益才[1] 

机构地区：[1]重庆交通大学信息科学与工程学院,重庆400074

出　　处：《计算机应用与软件》2016年第5期154-157,211,共5页Computer Applications and Software

基　　金：重庆市教委科学技术研究项目(KJ130421)

摘　　要：为从大量社会车辆中识别出疑似非法营运的车辆,提高交通管理部门行政执法的目的性和针对性,维护道路运输市场秩序,消除交通安全隐患。结合RFID车辆信息数据提出了基于k-mediods的非法营运车辆识别算法,并针对k-mediods算法缺点进行了基于距离贡献率和算法偶然性的2种改进。非法营运车辆识别的实现,首先需要提取出车辆RFID数据,并对其进行预处理,进而得到车辆运行行为数据,再利用PCA处理得到车辆运行特征数据,最后通过k-mediods算法聚类分析识别出非法营运车辆。实验结果表明,算法流程清晰,能够有效地识别出非法营运车辆。同时通过对算法进行改进,提高了算法稳定性和对非法营运车辆的正确识别数量,降低了错误识别数量。In order to recognise the vehicles to be suspected operating illegally from a large number of social vehicles,improve the purposive intention and pertinence of the traffic administration in administrative enforcement,maintain the market order of road transportation and eliminate the hidden traffic safety hazard,we proposed the k-mediods-based illegal operation vehicles recognition algorithm in combination with RFID vehicle information data. In light of the shortcomings of k-mediods algorithm,we made two improvements on it,the distance contribution rate-based and the algorithm contingency-based respectively. To implement the illegal operation vehicles recognition,firstly it needs to extract vehicles＇ RFID data,and carries out pretreatment on them,and then obtains the behaviour data of vehicles operation;secondly,it acquires the characteristics data of vehicles operation by means of PCA processing; finally,it recognises the illegal operation vehicles by k-mediods-based clustering analysis. Experimental results show that the algorithm has a clear flow,it can recognise the illegal operation vehicles effectively,meanwhile it can improve the stability of the algorithm by improving it,and in turn raises the numbers of correct recognition of the illegal operation vehicles,and decreases the numbers of wrong recognition as well.

关 键 词：k-mediods 距离贡献率 偶然性 RFID 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]