基于重要度计算的语义物联网本体摘要方法研究

发布时间：2018-02-09 01:35

本文关键词： 语义物联网本体摘要重要度计算 RDF语句链接分析　出处：《大连海事大学》2017年硕士论文　论文类型：学位论文

【摘要】：针对物联网的内在矛盾,将语义协同过程引入物联网,形成新一代网络,即语义物联网。语义协同过程是指基于本体进行语义标注与语义理解,因此,本体在语义物联网中起到了核心作用。但是,语义物联网本体目前主要存在两点不足:一是本体数量过多、内容复杂,不利于本体工程专家去理解并完成本体重用等相关工作;二是本体体积过大,对于语义物联网底层设备资源能力受限的情况下,能够提供一个精简且涵盖了最重要信息的本体是很有必要的。为了解决语义物联网本体不利于理解、体积过大等问题,本文提出了一种基于重要度计算的语义物联网本体摘要方法,从而保证语义物联网语义协同过程的更好实现。本论文采用结构和语用特征相结合的方式来实现基于重要度计算的语义物联网本体摘要。首先,对语义物联网本体进行预处理,获取本体中的术语集、RDF语句集和链接类型集,生成初始的RDF语句序列,采用RDF语句作为语义物联网本体摘要过程中的基本语义单元,创建表征本体结构的二分图模型;接着,在二分图模型的基础上,对RDF语句进行基于链接分析的结构重要度计算和基于语用统计的语用重要度计算,通过对统计结果和阈值的比较,进行RDF语句综合重要度计算,生成按照重要度排序的RDF语句序列,并设计了重要度计算过程算法;然后,对按照重要度排序的RDF语句序列进行冗余处理,将综合重要度最高的对象抽取到摘要结果集中,按需设定摘要长度后,生成摘要后结果集,并设计了摘要生成过程算法。为了验证上述方法在语义物联网中的有效性,本论文最后设计并实现了基于重要度计算的语义物联网本体摘要原型系统。该系统从功能上分为语义物联网本体预处理模块、语义物联网本体重要度计算模块及基于重要度的语义物联网本体摘要生成模块,分别实现了对语义物联网本体的解析、建模,重要度计算,消除冗余与摘要抽取功能。最终通过实验验证了该方法的准确性、可行性及该原型系统的性能。实验结果表明该本体摘要方法可以为用户提供满足其偏好的最优本体。
[Abstract]:In view of the inherent contradiction of the Internet of things, semantic collaborative process is introduced into the Internet of things to form a new generation of network, namely semantic Internet of things. Semantic collaborative process refers to semantic annotation and semantic understanding based on ontology. Ontology plays a central role in the semantic Internet of things. However, there are two main shortcomings in semantic ontology: first, the number of ontology is too much and the content is complex, which is not conducive to ontology engineering experts to understand and complete ontology reuse and other related work; Secondly, it is necessary to provide a concise ontology that covers the most important information in order to solve the problem that semantic Internet of things is not conducive to understanding, because the ontology is too large, and it is necessary for the semantic Internet of things to provide a concise ontology that covers the most important information under the condition of limited resource capacity of the underlying devices in the semantic Internet of things. In this paper, a semantic Internet of things ontology summary method based on importance calculation is proposed. In order to ensure the better realization of semantic collaborative process of semantic Internet of things. This paper uses the combination of structure and pragmatic features to realize semantic ontology summary based on importance calculation. Firstly, the semantic ontology of Internet of things is preprocessed. The RDF statement set and link type set are obtained, and the initial sequence of RDF statements is generated. The RDF statement is used as the basic semantic unit in the process of semantic ontology summary, and the bipartite graph model is created to represent the ontology structure. On the basis of bipartite graph model, we calculate the structural importance of RDF sentences based on link analysis and pragmatic statistics based on pragmatic statistics. By comparing the statistical results and threshold values, we calculate the comprehensive importance degree of RDF sentences. The sequence of RDF statements sorted according to the importance degree is generated, and the algorithm for calculating the importance degree is designed. Then, the sequence of RDF statements sorted according to the importance degree is processed by redundancy, and the objects with the highest synthetic importance are extracted into the summary result set. After the length of the summary is set according to the need, the result set is generated, and the algorithm of the process is designed. In order to verify the effectiveness of the above method in the semantic Internet of things, In the end of this thesis, the prototype system of semantic ontology abstracting based on importance computing is designed and implemented, which is divided into semantic ontology preprocessing module. The semantic Internet of things ontology importance calculation module and the semantic Internet of things ontology summary generation module based on the importance of the semantic Internet of things ontology analysis, modeling, importance calculation, respectively, Finally, the accuracy and feasibility of the method and the performance of the prototype system are verified by experiments. The experimental results show that the proposed method can provide users with the best ontology to satisfy their preferences.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.1;TP391.44;TN929.5

【参考文献】