基于Apriori算法的车辆检测相似重复记录消除方法  

Elimination Method for Approximately Duplicate Records in Vehicle Inspection Based on Apriori Algorithm

作　　者：安相璧[1,2] 杜艾永[2] 李树珉[1,2] 

机构地区：[1]天津大学精密仪器与光电子工程学院,天津300072 [2]军事交通学院汽车工程系,天津300161

出　　处：《天津大学学报》2010年第7期606-610,共5页Journal of Tianjin University(Science and Technology)

摘　　要：为消除在数据库中存在的中文相似重复记录,提出一种改进的Apriori算法,利用该算法获得数据库记录的频繁项集.基于频繁项集,消除进行比较记录的共有项,有效提高相异字符的计算权重.然后利用FRMA算法计算记录间的相似度,最终消除中文相似记录.在车辆检测数据库中对该算法进行了实验,取得了较好的实验结果,证明该算法具有较好的实用价值.In order to eliminate the Chinese duplicate records in database, an improved Apriori algorithm was put forward. Based on the group of frequent words obtained from the database entries with the proposed algorithm, the coexisting words in two records were eliminated and the weight of dissimilar characters was enhanced effectively. Then the similarity between the two records was calculated with FRMA algorithm so that the Chinese duplicate re- cords were eliminated finally. The algorithm was tested on the vehicle inspection database, and achieved satisfactory results which has proved the application value of the algorithm.

关 键 词：相似重复记录 APRIORI算法 FRMA算法 

分 类 号：U463.3[机械工程—车辆工程]