基于矩阵加权关联规则挖掘的伪相关反馈查询扩展

提出一种面向查询扩展的矩阵加权关联规则挖掘算法,给出与其相关的定理及其证明过程.该算法采用4种剪枝策略,挖掘效率得到极大提高.实验结果表明,其挖掘时间比原来的平均时间减少87.84%.针对现有查询扩展的缺陷,将矩阵加权关联规则挖掘技术应用于查询扩展,提出新的查询扩展模型和更合理的扩展词权重计算方法.在此基础上提出一种伪相关反馈查询扩展算法——基于矩阵加权关联规则挖掘的伪相关反馈查询扩展算法,该算法能够自动地从前列n 篇初检文档中挖掘与原查询相关的矩阵加权关联规则,构建规则库,从中提取与原查询相关的扩展词,实现查询扩展.实验结果表明,该算法的检索性能确实得到了很好的改善.与现有查询扩展算法相比,在相同的查全率水平级下,其平均查准率有了明显的提高.

Query Expansion of Pseudo Relevance Feedback Based on Matrix-Weighted Association Rules Mining

An algorithm of matrix-weighted association rule mining for query expansion is presented based on the quadruple pruning, and a related theorem and its proof are given. This method can tremendously nhance the mining efficiency. Experimental results demonstrate that its mining time is averagely reduced by 87.84%, compared to that of the original one. And a query expansion algorithm of pseudo relevance feedback is proposed based on matrix-weighted association rule mining, which combines the association rules mining technique with the query expansion. The algorithm can automatically mine those matrix-weighted association rules related to the original query in the top-ranked retrieved documents to construct an association rules-based database, and extract expansion terms related to the original query from the database for query expansion. At the same time, a new computing method for weights of expansion terms is given. It makes the weighted value of an expansion term more reasonable. Experimental results show that this method is better than traditional ones in average precision.