基于引用的科技文献排名算法

Citation-based Scientific Documents Ranking Algorithm

[目的/意义]旨在将科技文献的价值进行量化,提高PageRank算法应用在科技文献排名中的准确性。[方法/过程]在加入时间因子的PageRank算法的改进算法WPageRank的基础上,加入引用相关度进行改进,并计算文献的固有价值,与文献的PageRank值进行加权求和,得到文献的最终价值。[结果/结论]本文提出的方法使新发表的高质量文献也可以获得较高排名,并且使领域内的高质量文献更容易被检索到,同时保证了检索的时效性和主题集中性。

[Purpose/significance]The paper is to quantify value of scientific documents,and improve the accuracy of PageRank algorithm used in ranking of scientific documents. [Method/process]Based on improved algorithm of WPageRank which is a PageRank algorithm with adding time factor,the paper improves this algorithm with adding relevance of reference,calculates intrinsic value of documents,weighted sums with PageRank of documents,and obtains final value of documents. [Result/conclusion]The method proposed in this paper can make the newly published documents rank higher,and high-quality documents in a field easier to be retrieved. This method guarantees time effectiveness and topic concentration of the retrieval at the same time.