面向网页文本的地理信息变化语义检测方法研究

发布时间：2018-05-18 10:34

本文选题：地理信息 + 变化检测　；参考：《南京师范大学》2013年硕士论文

【摘要】：本论文依托国家测绘科技项目“网络地理信息变化检测技术研究”,面向地理信息动态持续更新需求,针对网络文本中地理信息变化语言的描述特点,集成地名识别、时间和属性抽取技术,较为系统地探索了面向网页文本的地理信息变化语义检测方法,研发了相应的原型系统,并进行了实验验证分析。主要研究内容和结论包括以下三个方面： (1)地理信息变化语义检测知识库构建：通过对地理信息变化网页中变化信息内容和相关网页获取方法的分析,构建了包括要素特征词汇库、要素变化词汇库、停用词词汇库、空间关系词汇库等在内的变化检测知识库,利用Protege本体编辑器和OWL API实现了知识库的管理和维护。 (2)地理信息变化网页获取与信息解析：设计了基于搜索引擎的主题爬虫和通用主题爬虫,有效解决了相关网页的广度和深度搜索问题；构建了网页可信度计算模型,为虚假网页的甄别提供了依据；在时间抽取和地名识别技术基础上,提出了知识库驱动的网页文本中地理信息变化特征信息抽取方法。 (3)原型系统研发和实验验证：在上述研究基础之上,借助富客户端技术设计了面向网页文本的地理信息变化语义检测原型系统,实现了知识库管理、变化信息获取和空间统计分析等功能,并以南京地区为例,进行了实验验证分析。研究表明,地理信息涉及地理要素类型较多,知识库构建是一个较为复杂的系统工程,采用人工归纳存在一定的局限性。不同地理要素类型的变化信息在网页文本中的出现频率差异较大,交通和水系类地理信息解析效果较好。另外,大多数从网页文本中获取的地理信息变化属性为定性描述,而且具有较强的模糊性和不确定性,需要结合其他地理信息数据源进行进一步验证。
[Abstract]:This paper relies on the national surveying and mapping project "the research of the network geographic information change detection technology", facing the geographical information dynamic continuous renewal demand, according to the description characteristic of the geographical information change language in the network text, integrates the place name recognition, Based on the technology of time and attribute extraction, this paper systematically explores the semantic detection method of geographic information change oriented to web page text, develops the corresponding prototype system, and carries out the experimental verification and analysis. The main research contents and conclusions include the following three aspects: 1) Construction of knowledge base for semantic detection of geographic information change: by analyzing the content of change information and the methods of obtaining relevant web pages, this paper constructs a lexical database including the feature of elements, the lexicon of change of elements, and the vocabulary of inactive words. The Protege ontology editor and OWL API are used to manage and maintain the knowledge base. (2) Web page acquisition and information analysis of geographic information change: a search engine based topic crawler and a general theme crawler are designed, which effectively solve the breadth and depth search problems of related web pages, and a web credibility calculation model is constructed. Based on the technology of time extraction and place name recognition, this paper proposes a method of extracting the feature information of geographical information change in web pages driven by knowledge base. Research and development of prototype system and experimental verification: on the basis of the above research, a prototype system for semantic detection of geographic information change oriented to web page text is designed with the help of rich client technology, and the knowledge base management is realized. Taking Nanjing area as an example, the functions of information acquisition and spatial statistical analysis are analyzed. The research shows that geographical information involves many types of geographical elements, and the construction of knowledge base is a more complex system engineering, and the artificial induction has some limitations. The frequency of changing information of different geographic element types in the text of web pages is quite different, and the interpretation effect of traffic and river type geographic information is better. In addition, most of the geographic information attributes obtained from the text of the web page are qualitative description, and have strong fuzziness and uncertainty, which need to be further verified with other geographic information data sources.
【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】：P208;TP391.1

【参考文献】