传统的 Top-k 查询处理都是利用单用户偏好来计算评分函数，这种方法有极大的局限性。针对基于多用户偏好的 Top-k 查询处理问题进行研究，为了提高查询效率，首先提出了预处理算法 PA 与 PVA ，生成一些具有代表性的系统用户偏好，并据此将初始数据集进行全排序，保存在物化视图中，以便利用它们进行 Top-k 查询。然后，提出了处理 Top-k 查询的 VBA 算法且进行了正确性与完备性论证。最后，实验结果表明，该算法比直接在原数据集中查询的效率有极大的提高。%The traditional Top-k query processing is to use a single user preference to calculate score function ,which has greatl imitations .The problem of Top-k query based on muti-user preferences is studied .In order to improve the efficien-cy of query processing ,the preprocessing algorithm PA and PVA is proposed to generate some typical user preferences as system user preferences and sort the initial data set according to these preferences and store the results on the materialized views ,which can be used to process Top-k query .Then ,the VBA algorithm dealing with Top-k queries and demonstrated the correctness and completeness of the algorithm is proposed .Finally ,the experimental results show that our algorithm has improved the query processing greatly compared to the nature algorithm which runs directly in the original data set .
展开▼
