一种基于主题爬行模式的地理信息分布式检索方法  

A Geographic Information Retrieval Methods Based on the Mode of Distributed Crawling

作　　者：王小康[1] 邓硕[1] 吴博[1] 李景文[1,2] 

机构地区：[1]桂林理工大学测绘地理信息学院,广西桂林541004 [2]桂林理工大学广西空间信息与测绘重点实验室,广西桂林541004

出　　处：《测绘与空间地理信息》2015年第4期96-97,101,共3页Geomatics & Spatial Information Technology

基　　金：广西自然科学基金重点项目(2014GXNSFDA118032)资助

摘　　要：当前网络中地理信息以几何形式递增,为了高效地从海量网络信息中检索出高质量的地理信息,本文提出了一种基于主题爬行的地理信息分布式检索方法。本文采用面向对象的方法将网络地理数据按照四元组的要求进行分解和组织,对地物实体的主题文本特征、地理空间特征、时间维特征等相关信息进行封装,建立四元组实体对象,实现了地理信息数据的相互集成与组织。引入MapReduce模式的并行处理机制完成对网页中地理信息数据的优化存储与索引,并且通过分别计算网页文本、地理文本与查询关键词的主题相关性对爬取的网页进行有序的排列,从而提供快捷、高效的地理信息主题查询。Current geographic information network in order to geometric form is increasing,in order to efficiently retrieve from the massive network information of high quality geographic information,this paper proposes a geographic information retrieval methods based on distributed crawling crawling body. In this paper,by using object oriented method the network geographic data are decomposed and organization in accordance four tuples of objects,,geographic features,time Victor syndrome and other related information package,the establishment of four tuple entity object,realization of the geographic information data integration and organization,enhance organizational efficiency index of the source data. Parallel processing mechanism into MapReduce mode to accomplish optimal storage of geographic information data in web and retrieval,and by calculating the page relevance text,geography text with the query keywords are orderly arranged for crawling web pages,so as to provide geographic information subject fast,efficient query.

关 键 词：MAPREDUCE 主题爬行 地理信息 主题相关度 

分 类 号：P208[天文地球—地图制图学与地理信息工程]