论述了信息检索中的向量空间模型、概率模型以及语言模型中所采用的相关反馈技术。其中主要介绍检索词的权重调整、查询扩展、文档相关反馈,以及语言模型中的查询语言模型和文档语言模型的调整。针对最近反馈方面的最新成果——基于term的反馈技术进行了探讨,指出了相关反馈在今后研究的方向,即提供个性化的如分层反馈和利用日志进行反馈,并讨论了相关反馈技术对检索性能的影响。
This paper introduced relevance feedback mothod about vector space model, probability model and language model in information retrieval fields, elaborated how to regulate the weights of terms, expands query, relevance feedback of documents, regulate about the query language model and document language model. It also introduced how to adopt relevance feedback to improve information retrieval performance. It discussed the novel technology of feedback with term feedback technology. At last, discussed the problems of the future direction in relevance feedback, the personality of hierachical feedback and the feedback of using log .