基于本体的个性化信息检索研究

发布时间：2018-05-25 21:54

本文选题：本体 + 个性化检索　；参考：《兰州理工大学》2013年硕士论文

【摘要】：随着互联网技术的快速发展,网络信息资源呈现出一种爆炸式的增长态势,与此同时增加的还有人们的信息需求量,但是要从这个庞大的信息海洋中快速并准确地定位到自己所需要的信息,却比较困难,进而导致了所谓的“信息过载”与“信息迷向”问题。其中,最主要的原因是通用搜索引擎没有将用户的个性化需求纳入到检索过程中,提供千篇一律的检索结果。因此,如何了解用户意图,掌握用户的个性化信息,并提供个性化服务成为信息检索领域中许多学者的研究热点。本文围绕着基于本体的个性化检索引擎的研究逐步展开,针对个性化信息检索的关键技术——用户兴趣模型的构建与更新机制进行了深入剖析,并在原有语义相似度算法基础上进行了改进。本文研究的主要内容如下： (1)将软件工程学科的思想引入到领域本体的构建过程,参照Customer本体库与领域本体的构建准则、构建方法构建了领域本体。 (2)针对现有信息检索系统模型的特点,分析个性化信息检索的需求,设计了基于本体的个性化信息检索的原型,并且详细介绍了各模块的功能。重点论述了信息检索中语义相似度算法的改进算法,并且将其应用于个性化信息检索技术中,提高了检索性能。 (3)利用本体思想与分析用户浏览行为,构造了多层分级的用户兴趣模型。通过利用用户兴趣漂移思想,加入遗忘因子,动态捕捉用户兴趣的更新,建立了用户兴趣模型的更新遗忘机制。通过实验的分析验证,构造的用户兴趣模型能够较为准确的描述用户兴趣的改变,从而更加准确的提供信息检索功能。
[Abstract]:With the rapid development of Internet technology, network information resources show an explosive growth trend, at the same time, there is an increase in people's demand for information. However, it is difficult to locate the information needed quickly and accurately from this huge ocean of information, which leads to the so-called "information overload" and "information obsessive" problems. The main reason is that the general search engine does not incorporate the user's personalized requirements into the retrieval process and provide uniform retrieval results. Therefore, how to understand the user's intention, grasp the user's personalized information, and provide personalized services has become a hot research topic for many scholars in the field of information retrieval. In this paper, the research on the ontology based personalized retrieval engine is carried out step by step. The key technology of personalized information retrieval is the construction and updating mechanism of user interest model. The algorithm is improved on the basis of the original semantic similarity algorithm. The main contents of this paper are as follows: 1) the idea of software engineering is introduced into the construction process of domain ontology. According to the construction criteria of Customer ontology library and domain ontology, the construction method of domain ontology is constructed. According to the characteristics of the existing information retrieval system model, the requirements of personalized information retrieval are analyzed, the prototype of personalized information retrieval based on ontology is designed, and the functions of each module are introduced in detail. This paper mainly discusses the improved algorithm of semantic similarity algorithm in information retrieval, and applies it to personalized information retrieval technology to improve the retrieval performance. 3) using ontology theory and analyzing user browsing behavior, a multi-layer hierarchical user interest model is constructed. By using the idea of user interest drift, adding the forgetting factor to capture the update of user interest dynamically, the updating forgetting mechanism of user interest model is established. Through the analysis and verification of experiments, the constructed user interest model can accurately describe the change of user interest, thus providing more accurate information retrieval function.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.1

【参考文献】