基于信息资源不同一性的检索结果优化排序

Retrieval Results Re-ranking Based on the Nonidentity of Information Resource

相关反馈一直是优化检索结果的主要方法,可归纳为四种模式。不同一性是信息资源的重要特性,用户的个性化信息是进行相关反馈的重要参考;利用用户的个性化检索信息进行相关反馈,并按照不同一性对反馈结果进行排序是一种全新的检索结果优化排序方法,该方法兼有更准确表达用户需求、个性化排序、提高用户对相关反馈检索结果满意度、容易实现等多种优势,具有较高的理论价值和应用前景。

One main method for query expansion is relevance feedback, which has four modes. Nonidenti- ty is an important property of information resource, and the user＇ s personalized retrieval information is the significant basis for relevance feedback. Therefore, a new retrieval results re-ranking method which used the user＇ s personalized retrieval information to the relevance feedback and re-rank the results based on the nonidentity of information resource is proposed. plication prospect can convey the use＇s information The method has high theoretical value and broad ap- need more explicitly, personalize the ranking results, improve the use＇ s satisfaction to the relevance feedback,moreover, be realized easily.