中文摘要：

墒要：将时态信息融入到信息检索技术中可以有效提高检索效果。时态信息检索已有较多的研究，而现有数据库信息检索方法还缺乏对时态信息的有效利用。针对这一研究问题，提出关系数据库上基于时态语义的关键词检索方法，引入时态信息构建时态数据图，设计时态相关性评分机制，在时态图搜索过程中引入时态语义约束，设计基于关键词的时态检索算法。实验验证了该方法可以有效提高数据库信息检索效果，而检索性能并没有降低。

英文摘要：

Temporal information retrieval （TIR） can effectively improve the effect of retrieval results, and has been widely studied in information retrieval community. However, existing database information retrieval methods lacked of effective utiliza- tion of temporal information. For this research problem, this paper proposed a keyword-based TIR method over relational data bases. It built temporal data graph and designed temporal correlation scoring mechanism, then employed temporal semantic constraints in the process of temporal graph search. Experiments show that this method can obviously improve the effect of the data-base information retrieval, and the retrieval performance is acceptable.