QWS-Rank:一种新颖的XML关键词搜索结果排序方法

QWS-Rank: a Novel Ranking Method for XML Keyword Search

提出一种新颖的XML关键词搜索结果排序方法 QWS-Rank.这种方法基于一种新的XML关键词查询模型QWS,将一个关键词查询理解为由若干查询单元组成的带结构的查询.这种排序方法从关键词查询的结构出发,根据查询结果与关键词查询中的查询单元匹配的数量和质量(即匹配程度)对查询结果计分,并考虑了查询结果的结构,使得这种排序方法较好地揭示了查询结果与查询的语义相关性.实际数据集上的实验结果显示,提出的方法产生的结果要优于现有方法.

This paper proposes a novel ranking method, called QWS-Rank, for XML keyword search. The method is based on a new XML keyword query model, called QWS, which interprets a keyword query as a query-with-structure consisting of several query units. The ranking method puts emphasis on the query structure, and scores the query results based on how many and how well the query result matches the query units. It also takes the query structure into consideration. The merit of this method is that it reveals the semantic relevance between the query result and the query. Experimental results on real datasets show that the proposed method ourperforms previous one.