Effectively extracting useful information from complex structured XML databases

In this paper, we focus on providing users with an effective and efficient way to extract useful information from complex structured XML databases. We propose a new query mechanism to provide users with the ability to refine the ambiguous semantics of the initiative ATP query before execution, thus we can avoid identifying meaningful answers from large amount of irrelevant results returned by existing methods. We implement algorithms for efficient evaluating an ATP query. Rich experimental results show that our method can be used easily and outperform existing query engines significantly.