基于关键词的关系数据库时态信息检索方法研究

发布时间：2017-12-27 08:25

本文关键词：基于关键词的关系数据库时态信息检索方法研究　出处：《大连海事大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着大数据时代的来临,大量积累的数据已成为了各行各业重要的数据资产。时间是体现数据价值的重要指标,时态数据越来越得到人们的关注。因此如何对时态数据进行有效的存储、管理和检索已成为数据库和信息检索领域研究的热点。数据库信息检索技术是数据库和信息检索的交叉领域,可以有效的支持普通用户在数据库上高效地进行关键词检索,该领域已取得了众多的研究成果。时态信息检索领域研究表明,通过将时态信息融入信息检索技术,可以有效地处理用户的时态查询,快速、高效地检索用户所需要的信息。然而现有的关系数据库关键词检索方法未考虑数据时态性,缺乏对时态数据的检索。因此针对这一问题,本文从时间维度出发,研究关系数据库上基于关键词的时态信息检索方法。首先,介绍时态信息处理的相关理论,为本文时态检索方法的研究提供思路和理论支持。然后,在原有关系数据库关键词检索方法基础上,引入时间维度提出了时态信息检索模型,设计了基于关键词的关系数据库时态信息检索方法。该方法包括三部分内容:(1)通过分析数据库中存储的时态实体和实体之间的时态关系,构建时态数据图;(2)由于现有的索引方法不能满足时态关键词结点的快速查找,设计了时态倒排索引,该索引通过对每个关键词对应的时态结点集合进行时态分区来提高查找效率;(3)设计了时态检索算法T-STAR,该算法主要采用时间修剪策略,对不满足时态约束的边进行修剪,保证检索结果满足时态查询的时间约束,同时,提出了时态边权重的计算方法,更好的满足检索结果的内容相关性。最后,实现了基于关键词的关系数据库时态检索原型系统,利用Employees和NBA时态数据集从P@K和MAP两个评价指标对本文提出的方法进行了实验评估。实验结果表明了该方法在保证检索效率的前提下,有效提高了数据库信息检索效果,满足用户的时态检索需求。
[Abstract]:With the advent of the large data age, a large amount of accumulated data has become an important data asset in all walks of life. Time is an important indicator of the value of data, and more and more attention has been paid to the temporal data. Therefore, how to effectively store, manage and retrieve temporal data has become a hot topic in the field of database and information retrieval. Database information retrieval technology is a cross field between database and information retrieval. It can effectively support ordinary users to search keywords efficiently in database, and has achieved many research results in this field. In the field of temporal information retrieval, it is shown that temporal information can be applied to information retrieval technology, which can effectively handle temporal queries of users, and retrieve users' information quickly and efficiently. However, the existing relational database keyword retrieval methods do not take into account the temporal nature of the data, and lack the retrieval of temporal data. Therefore, in view of this problem, this paper studies the temporal information retrieval method based on key words on the relational database from the time dimension. First, the relevant theory of temporal information processing is introduced, which provides ideas and theoretical support for the research of the temporal retrieval method. Then, based on the keyword search method of original relational database, a temporal information retrieval model is proposed by introducing time dimension, and a relational database temporal information retrieval method based on keywords is designed. The method includes three parts: (1) the temporal relationship between the temporal entity and stored in the database analysis, the construction of temporal data; (2) due to the existing indexing methods to meet the fast search keywords temporal node, design of temporal inverted index, the index of temporal partition set to improve the search efficiency the temporal node for each keyword corresponding; (3) the design of temporal retrieval algorithm T-STAR, the algorithm mainly uses the pruning strategy to prune the time, does not meet the temporal constraints, ensure the retrieval results meet the time constraints, temporal query at the same time, put forward the method of calculating weights of temporal edge content relevance to better meet the needs of retrieval the results of the. Finally, a relational database temporal retrieval prototype system based on keywords is implemented. Based on Employees and NBA temporal data set, two evaluation indexes from P@K and MAP are used to evaluate the method proposed in this paper. The experimental results show that the method can effectively improve the database information retrieval effect and meet the user's temporal retrieval requirements under the premise of ensuring the retrieval efficiency.
【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】：TP311.13;TP391.3

【参考文献】