分析了现有搜索引擎重排序方法的不足，并在此基础上提出了基于文档间距离的重排序算法。利用相关性函数，表示出文档问的相关性与文档间距离的关系，通过计算得到文档之间距离的大小顺序，进而得到文档的相关性大小顺序，然后根据文档所属主题概率，最终得到文档的排序。实验表明，该算法大大提高了用户的检索速度与质量，优化了搜索引擎的性能。%This paper analyzes the shortages of the existing re -ranking methods of the search engine, gives a theoretical description of the re - ranking algorithm based on the inter - document distances. The relationship of the documents' relevance and documents' distances is expressed by the relevance function, and the rank of in- ter- document distance can be eaeulated, then the rank of the documents' relevance are obtained. Finally, the rank of document obtained according to the topic probability belonged to documents. The experiments shows that, the algorithm greatly improves the user's retrieval speed and quality, optimizes the performance of search engine.
展开▼
