地理信息公共服务平台地名信息检索方法研究

发布时间：2018-06-15 01:54

本文选题：地理信息 + 公共服务平台　；参考：《南京师范大学》2013年硕士论文

【摘要】：地名是人们赋予某一特定区域范围的地理实体的一种语言文字代号,同时也是用以区别另一特定地理实体的一种语言文字标识。地名信息是最常用的一类公共信息,不仅与人们的日常生活息息相关,而且是国家行政管理,经济建设,国内外交往不可或缺的基础信息资源。由于地名信息既包含了文本的属性信息,也蕴涵了丰富的时空信息,因此成为了连接空间信息与非空间信息的重要媒介,并在国内“数字城市”的地理信息公共服务平台的建设中发挥着日益重要的作用。随着地理信息公共服务平台建设的逐步深入及应用规模的急剧加速,平台中与地名相关的各类地理信息资源类型多样且日益庞杂。针对当前公共服务平台普遍利用传统空间查询查找地名所面临的检索效率低下、匹配精度不高、检索结果信息量有限的方法问题,本文在深入分析地理信息公共服务平台地名分类和传统地名查找方法缺陷问题的基础上,提出了基于领域特征词和Lucene的地理信息公共服务平台地名信息检索方法,并在对浙江省地理空间数据交换和共享平台的数据实验基础上,对方法进行了检索质量和效率的验证。本文主要的研究内容和成果包括以下几个方面： (1)地名领域特征词与索引的管理方法研究针对本文提出的平台多源地名的含义与分类体系,研究基于信息熵的高频词统计方法,结合人工筛选,构建领域特征词库及相应管理方法；研究基于Lucene的平台多级地名全文索引构建、维护方法。 (2)基于Lucene和领域特征词的地名检索与排序方法研究研究基于Lucene和领域特征词的地名检索方法实现的技术框架。利用特征词库进行分词解析,研究面向特征词库的查询表达式解析方法；研究根据文本相关性和地理相关性构成综合评价指标对地名信息检索结果进行排序优化。 (3)地名信息检索原型设计和方法验证以浙江省地理空间数据交换和共享平台中的地理空间数据作为实验对象,设计基于本文提出的领域特征词和Lucene的地名信息检索原型系统。通过SQL模糊查询和Oracle全文索引两种在地理信息公共服务平台中常用的地名检索方法进行对比分析,验证本文提出的地名信息检索方法的有效性。
[Abstract]:The place name is a kind of language and character code that people assign to a geographical entity in a certain area, and it is also a kind of language and character mark used to distinguish another geographical entity. Toponymic information is the most commonly used kind of public information, which is not only closely related to people's daily life, but also an indispensable basic information resource for national administration, economic construction and communication at home and abroad. Because the toponymic information not only contains the attribute information of the text, but also contains abundant space-time information, it becomes an important medium to connect the spatial information with the non-spatial information. And it plays an increasingly important role in the construction of the public service platform of geographic information in the domestic "digital city". With the gradual deepening of the construction of geographical information public service platform and the rapid acceleration of its application scale, all kinds of geographical information resources related to geographical names in the platform are diverse and increasingly complex. In view of the problems of low retrieval efficiency, low matching accuracy and limited amount of information in the search results, the public service platform generally uses traditional spatial query to find place names, which is characterized by low retrieval efficiency, low matching accuracy, and limited information content. Based on the in-depth analysis of geographical names classification and traditional toponymic search methods of geographical information public service platform, this paper puts forward a geographical information retrieval method based on domain feature words and Lucene. The retrieval quality and efficiency of the method are verified on the basis of the data exchange and sharing platform in Zhejiang Province. The main contents and achievements of this paper include the following aspects: 1) the research on the management method of feature words and indexes in the field of geographical names the meaning and classification system of the platform multi-source place names put forward in this paper. This paper studies the statistical method of high-frequency words based on information entropy, constructs domain feature thesaurus and corresponding management method, and studies the full-text index construction of multi-level geographical names based on Lucene platform. Maintenance method. Research on the method of place name Retrieval and sorting based on Lucene and Domain feature words; the technical framework of the realization of geographical name retrieval method based on Lucene and domain feature words. Using the feature lexicon to parse the word segmentation, the query expression parsing method for the feature dictionary is studied. This paper studies the ranking and optimization of toponymic information retrieval results according to the comprehensive evaluation index of text relevance and geographical relevance. The prototype Design and method Verification of toponymic Information Retrieval in Zhejiang Province Geospatial data in a spatial data exchange and sharing platform is used as an experimental object. A toponymic information retrieval prototype system based on domain feature and Lucene is designed. Through the comparison and analysis of two common geographical names retrieval methods, SQL fuzzy query and Oracle full-text index, the effectiveness of the toponymic information retrieval method proposed in this paper is verified.
【学位授予单位】：南京师范大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P281;P208

【参考文献】