模糊集方法在检索评价系统中的应用

评价一个信息检索系统的标准是该系统在多大程度上满足用户的能力。文中从传统的评价信息检索系统的两大标准———计算查全率和查准率的公式出发,结合实际情况,考虑检索结果相关性是一个模糊的概念,为了对这个模糊的概念以客观的度量,应用模糊集的方法对传统的查全率和查准率的公式进行了两种修改。一种修改是在每篇文档对总的查全率和查准率的影响相同时进行的,而第二种修改则是在每一类文档对总的查全率和查准率的影响相同时进行的。这两种修改是对传统的基于二值相关性的检索评价系统的有效扩充。

Application of Fuzzy Measure in Information Retrieval Evaluation

The criterion of evaluating an information retrieval system is the ability of how much the system contents the needs of users.Based onthe traditional formula in computing recall and precision,the paper considers the fuzzy property of relevance.We considerate that the retrieval results are fuzzy.In order to measure the fuzzy thing objectively,use fuzzy measure to revise the formula of traditional recall and precision in two ways.One is working at the time that every text has the same effect to the recall and precision,the other is working at the time that every class has the same effect to the recall and precision.These revisions efficiently widen the traditional retrieval evaluation which based on binary relevance.