关系数据库关键词查询处理关键技术研究

发布时间：2018-03-13 05:03

本文选题：关系数据库　切入点：关键词查询　出处：《黑龙江大学》2012年硕士论文　论文类型：学位论文

【摘要】：随着数据库技术的高速发展，普通用户可以像网页搜索引擎一样通过输入关键词直接在关系数据上查询结果，而不需要了解底层的数据库模式以及复杂的SQL查询语句。当前的关系数据库关键词查询系统还有一些不足的地方，例如，查询系统的效率不高，查询准确率低，不支持短语查询，查询的结果展示比较单一等。本文针对上述的不足做了一些改进工作。第一，提出了CNI方法来提高候选网的生成速度。原有的基于模式图查询系统是在产生元组集图后再进行广度优先搜索，比较费时，研究发现针对OR语义的查询可以通过建立候选网索引的方法，，直接通过候选网索引得到所需要的候选网络。第二，提出了P-B方法来有效地识别关系数据库中存在的短语，并将识别出的短语反馈给用户，用户自主选择符合查询条件的短语，从而提高了查询性能和查询准确率。第三，提出了TCD方法来有效地对查询的结果连接树聚类，将相同语义的连接树分成一组，并给出每个聚类的描述，方便用户理解查询结果的语义，并快速找到准确的查询结果。理论分析和实验表明上述提出的三种方法是有效的。
[Abstract]:With the rapid development of database technology, ordinary users can like web search engines like the input keywords directly results in relational data, without the need to understand the underlying database schema and complex SQL queries. Keywords current relational database query system has some deficiencies, for example, the efficiency of query system is not high. The low accuracy of the query, does not support the phrase query, the query results show is single. Aiming at the above problems do some improvement work.
First, we propose the CNI method to improve the production rate of candidate networks. The original query system based on schema graph is generated in the tuple set graph and breadth first search, time-consuming, the study found that in view of the OR semantic query through the method to establish the candidate net index, directly from the candidate network required by the candidate network index.
Second, P-B method is proposed to identify the phrases existing in relational database effectively, and feedback the identified phrases to users. Users can choose the phrases that meet the query conditions independently, thus improving query performance and query accuracy.
Third, the TCD method is proposed to effectively cluster the query results, and divide the connection tree of the same semantic into a set, and give the description of each cluster, so that users can understand the semantics of query results and find the exact query results quickly.
Theoretical analysis and experiments show that the three methods proposed above are effective.

【学位授予单位】：黑龙江大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP311.13

【引证文献】