摘要： 在传统的信息检索方法中,对检索返回结果的处理较少,导致检索的精确率不高。针对该问题,讨论了基于熵原理对大量的召回结果进行加工处理,设计了无指导文本选择算法,去除表面相关而本体不相关或相关度不高的数据,从而大大提高检索的精确率。

关键词: 信息检索, 熵原理, 特征项

Abstract: Traditional methods of IR dispose indexing results so less that the precision of IR is not high enough. In order to solve this issue, this paper designs an algorithm of no leading text choice according to entropy theory. The algorithm can dispose a great deal of indexing results and wipe off the data relevant in surface but non-relevant in essence or relevant less. The experimental results based on this algorithm show that the precision of IR can be improved remarkably.

中图分类号: 

• TP391.3