CAJ | 학술논문

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。
당전망락중지리신식이궤하형식체증，위료고효지종해량망락신식중검색출고질량적지리신식，본문제출료일충기우주제파행적지리신식분포식검색방법。본문채용면향대상적방법장망락지리수거안조사원조적요구진행분해화조직，대지물실체적주제문본특정、지리공간특정、시간유특정등상관신식진행봉장，건립사원조실체대상，실현료지리신식수거적상호집성여조직。인입MapReduce모식적병행처리궤제완성대망혈중지리신식수거적우화존저여색인，병차통과분별계산망혈문본、지리문본여사순관건사적주제상관성대파취적망혈진행유서적배렬，종이제공쾌첩、고효적지리신식주제사순。
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 decom-posed 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, en-hance 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.