A Flexible XML Query and Results Ranking Approach

摘要

Users often have imprecise ideas when searching the XML data and may not be able to formulate queries that accurately express their query intentions. Therefore, the query predicates user submitted could not act as rigid constraints for the answer elements. This paper presents a novel approach, XFQR (XML Flexible Query & Ranking), for relaxing the original query to provide approximate answer elements to the user. XFQR extends the categorical query predicates with the most similar values by evaluating the similarity of different pairs of values in the query history and expands the numerical query range to nearby values by using the kernel density estimation technology. XFQR speculates the importance of each specified leaf node based on the user query and assigns the score of each value of leaf node according to its "desirableness" to the user, and then the relevant answer elements are ranked according to their satisfaction degree to the user's needs. The results of preliminary experiment have identified the efficiency of our XFQR approach.