地名数据语义一致性匹配研究

发布时间：2018-03-16 08:50

本文选题：空间数据　切入点：地名　出处：《西南石油大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网+时代的到来,全球信息化进程不断加快,地理信息系统迅速发展并被广泛应用,积累了大量的信息资源。在空间数据产生与更新的过程中,空间数据的一致性匹配决定着地理空间信息的可用性与有效性,日益受到广泛关注。其中,地名作为基础地理信息数据和地理空间数据的重要构成部分,其一致性匹配也成为了多源地理空间数据集成以及地理信息检索中的关键研究。同时,伴随人们对地理空间信息交流的需求持续增多,地名的语义已逐渐变成了地理信息沟通与表达以及地理服务的核心。地名数据的语义一致性匹配计算作为重要的理论问题,在消除地名歧义、判别重复地名和实现地名精准检索中都有重要的应用,有必要对其进行深入的研究。本文首先阐述了地名语义一致性匹配的研究背景与意义以及空间数据语义一致性匹配与地名一致性匹配的国内外研究现状,同时剖析了地名及一致性的概念与相关理论,探究了地名分类、地名特征、地名语义以及其特殊的构词方式,并对地名变化的原因与造成地名语义不一致的多种因素进行了分析。其次,本文针对已有的地名匹配方法中存在的不足,结合规范汉语地名的词语特点,模拟人们的认知习惯,将地名语义一致性的匹配与地名语义相似度的计算相联系,在综合地名通名与专名的语义相似度匹配算法的基础上,通过动态确定权重的方式对相似地名进行了综合语义相似度的计算,以达到地名语义匹配的目的。在通名语义相似度计算时,根据地名的分类建立了地名类型本体作为地名通名语义一致性计算时的载体,并总结阐述了已有的基于本体计算概念相似度的算法,选择了基于语义距离的算法来度量地名通名的相似度,且引入了多种影响因子对原有算法进行了优化改进。最后,对改进算法进行了试验分析。结合地名数据,通过C#程序语言编写,实现了地名语义匹配综合相似度计算平台,利用改进算法对地名通名语义相似度和地名综合相似度进行了计算,得到匹配结果,与人们的主观判断值和其他地名匹配算法进行了对比。结果表明算法贴近人们的认知,是可行且有效的。
[Abstract]:With the advent of the Internet era, the process of global information technology is speeding up, and the geographic information system (GIS) is rapidly developing and widely used, which has accumulated a large number of information resources. In the process of spatial data generation and updating, The consistency matching of spatial data determines the availability and validity of geospatial information, and has attracted more and more attention. Among them, geographical names are important components of basic geographic information data and geospatial data. The consistency matching has also become the key research in multi-source geospatial data integration and geographic information retrieval. At the same time, with the increasing demand for geospatial information exchange, The semantics of geographical names has gradually become the core of geographical information communication and expression and geographical services. As an important theoretical problem, the semantic consistency matching calculation of geographical names data is used to eliminate the ambiguity of geographical names. There are important applications in identifying repeated geographical names and in achieving accurate retrieval of geographical names. It is necessary to study it deeply. Firstly, this paper expounds the background and significance of the semantic consistency matching of geographical names and the research status of semantic consistency matching of spatial data and the matching of geographical names at home and abroad. At the same time, this paper analyzes the concept and related theory of place name and its consistency, and probes into the classification of geographical names, the characteristics of place names, the semantics of place names and their special word-formation methods. The reasons for the change of geographical names and many factors that cause the semantic inconsistency of geographical names are analyzed. Secondly, aiming at the shortcomings of the existing toponymic matching methods, this paper simulates people's cognitive habits by combining with the characteristics of the words and expressions of standardizing Chinese place names. Combining the matching of semantic consistency of geographical names with the calculation of semantic similarity of geographical names, on the basis of synthesizing the semantic similarity matching algorithm of general and proper names of geographical names, In order to achieve the purpose of semantic matching of geographical names, the synthetic semantic similarity of similar place names is calculated by dynamically determining the weights. According to the classification of geographical names, the toponym type ontology is established as the carrier of the semantic consistency calculation of geographical names, and the existing similarity algorithms based on ontology computing concepts are summarized and expounded. The algorithm based on semantic distance is selected to measure the similarity of general names of geographical names, and a variety of influence factors are introduced to optimize and improve the original algorithm. Finally, the improved algorithm is tested and analyzed. Through C # programming language, the Synthetical similarity calculation platform of place name semantic matching is realized, and the general name semantic similarity and the place name synthesis similarity are calculated by using the improved algorithm, and the matching results are obtained. The results show that the algorithm is close to people's cognition and is feasible and effective.
【学位授予单位】：西南石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P281;P208

【参考文献】