中文摘要：

提出一种将关键词查询转换为SPARQL查询的方法来进行RDF数据的搜索.首先,根据RDF本身的关联特点,构建一个压缩实体摘要图;然后,借助关键词与所在实体的索引,将所查询的关键词在该摘要图上进行定位,通过图双向搜索算法找出包含关键词实体的前k子图,获得查询实体之间的关系,再联合最初的关键词及他们的属性,构建SPARQL查询;最后使用SPARQL搜索引擎执行查询.实验结果表明,所提方法较其他方法有更快的响应时间及更高的准确率.

英文摘要：

A method of translating keyword queries to SPARQL queries was presented to implement RDF（ resource description framework） keyword search. Firstly,a condensed entity summary was constructed according to connections of RDF data. Then,keywords were located on the designated nodes of the summary graph by the OPS（ object predicate subject） index. Top-k subgraphs connecting all keyword entities would be found by a bidirectional search algorithm.Finally,SPARQL queries were obtained by incorporating inter-entity relationships of top-k subgraphs,keywords and their properties,and SPARQL queries were executed by a SPARQL search engine. The experimental results showthat a faster response time and a higher accuracy than the existing ones are achieved.