广西农业信息地理匹配引擎设计与实现  

Design and implementation of geographic matching engine for Guangxi agricultural information

作　　者：朱明[1,2] 何永宁 吴博 ZHU Ming;HE Yong-ning;WU Bo(School of Earth Science and Resources,China University of Geosciences,Beijing 100083,China;Geographic Information Center of Guangxi,Nanning 530023,China)

机构地区：[1]中国地质大学(北京)地球科学与资源学院,北京100083 [2]广西壮族自治区基础地理信息中心,南宁530023

出　　处：《南方农业学报》2019年第1期201-207,共7页Journal of Southern Agriculture

基　　金：广西创新驱动发展专项项目(桂科AA18118048)

摘　　要：【目的】研究高并发、大流量农业信息地理匹配引擎,改进其算法,解决广西区内壮语地名匹配问题,实现农业信息的自动匹配与空间定位,以满足农业大数据平台高并发、大流量的地理匹配需求。。【方法】通过改造开源的Solr全文搜索引擎,结合广西地名中的少数民族语言特点,扩充地名词典、设计数据组织方式与逆向分词算法、改进TF-IDF算法。【结果】在改进方法的基础上设计并实现了农业地理信息地理匹配引擎。经过第三方15484条数据测试,能够准确切分壮语地名,引擎在500并发下仍具有良好的响应速度,匹配准确率达98.43%。地理匹配引擎目前已应用到糖业发展大数据平台中,并取得了良好的效果。【建议】针对测试中出现的问题,建议在下一步工作中扩充并完善词库内容、增强语义推理能力、研究基于空间语义的定位算法,提高广西农业信息的定位精度。【Objective】This paper mainly studied and developed a geographic matching engine for agricultural information with high concurrency and large data flow. Through improving place name segmentation,searching and matching algorithms,problems of Zhuang language place name matching were resolved,which enabled agricultural information automatch with spatial localization,and met the geographical matching demand of high-concurrency and large data flow in agricultural big data platform.【Method】By reforming the Solr full-text search engine,a novel geographic matching engine was designed and implemented through absorbing characteristics of minority languages in Guangxi place names,expanding the geographical name dictionary,designing reverse word segmentation algorithm and improving TF-IDF algorithm.【Result】The agricultural geographic matching engine was developed based on the improved method. More than 15484 third-party entries were tested. The results showed that Zhuang place names could be divided accurately. The response speed of the engine was fast under 500 concurrency with accuracy of 98.43%. The engine has been applied in Sugarcane Industry Development Big Data Platform and achieved sound effects.【Suggestion】Based on the problems in the test,the experiment suggested to expand and improve the lexicon content,enhance the semantic reasoning ability,study the location algorithm based on spatial semantics,and improve the location accuracy of Guangxi agricultural information in the next step.

关 键 词：农业信息 地理匹配引擎 地名分词 地名检索 地名匹配算法 广西 

分 类 号：S126[农业科学—农业基础科学] P208[天文地球—地图制图学与地理信息工程]