关键词: Web数据库，不精确查询，关联规则，语义相似度

Abstract: To deal with the problem of answering the Web database imprecise queries, this paper proposed a semantic similarity-based Web database imprecise query approach. For a given query, one or several similar queries in the query history will be found firstly, and the similarity of each similar query to the original query is greater than the given rclax}r tion threshold. Then, the tuples matched to these queries arc treated as the imprecise query results to the current query. Finally, the result tuples are ranked according to their satisfaction to the original ctuery. Results of experiments de- monstrate that the query similarity measuring method proposed is stable and reasonable, and the imprecise query method proposed has higher recall and the ranking accuracy as well.

Key words: Web database,Imprecise query,Association rule,Semantic similarity