K叉树地址的模糊匹配研究与实现  

Research and Implementation of Fuzzy Matching for K-tree Address

作　　者：李新放[1,2] 宋转玲 陈学业[3] 贺彪 刘海行 LI Xinfang;SONG Zhuanling;CHEN Xueye;HE Biao;LIU Haixing(The First Institute of Oceanography,State Oceanic Adtninistration,Qingdao 266061,China;Laboratory for Regional Oceanography and Numerical Modeling,Qingdao National Laboratory-for Marine Science and Technology,Qingdao 266237,China;Shenzhen Research Center of Digital City Engineering,Shenzhen 518040,China)

机构地区：[1]国家海洋局第一海洋研究所,山东青岛266061 [2]青岛海洋科学与技术国家实验室区域海洋动力学与数值模拟功能实验室,山东青岛266237 [3]深圳市数字城市工程研究中心,广东深圳518040

出　　处：《测绘通报》2018年第9期126-129,155,共5页Bulletin of Surveying and Mapping

基　　金：国家重点研发计划(2016YFA0602200);中央级公益性科研院所基本科研业务费专项资金(2015G18; 2015P12)

摘　　要：在数字城市信息资源的集成和融合中,地名地址匹配是一项非常关键的基础技术。由于中文语义和地名地址描述的复杂性,中文地址的匹配比英文要复杂得多,基于海量中文地址数据进行准确分词,实现快速高效的地址匹配是城市数据集成融合的关键问题。本文在对现有地址编码及分词技术研究的基础上,通过一种基于规则和统计的组合方法来实现中文地址分词,并且使用K叉树的结构实现对中文地址的存储,提高了中文地址匹配查询的准确度和效率。基于预处理后的10 000个深圳市地址数据,通过开发原型系统对该方法进行了比较测试,验证了该方法的有效性。In the integration of digital city information resources, the address matching is a very cnacial basic technology. Due to the complexity of description of Chinese semantics and address,the matching of Chinese addresses is much more complicated than that of English.How to accurately segment words based on mass data of Chinese address and realize fast and efficient address matching is an urgent problem to be solved.Based on the research of existing address coding and word segmentation technology,this paper proposes a combination method based on rules and statistics to implement Chinese address segmentation, and uses K-tree to store the Chinese address and improves the Chinese address matching query Accuracy and efficiency.The method was tested by the prototype system based on 10,000 address data after pretreatment in Shenzhen City to verify the effectiveness of the method.

关 键 词：地址匹配 分词 模糊匹配 K叉树 

分 类 号：P208[天文地球—地图制图学与地理信息工程]