基于改进PageRank算法的引文文献排序方法

摘要

Based on the structural characteristics of the citation network, this paper identifies the value of scientific literatures effectively and establishes certain ranking relationships, which not only provides valuable references for scientific evaluation, but also enriches the scientific evaluation method. In view of the fact that the results obtained by the classical PageRank algorithm from the citation network tend to be the literatures published long ago, which is unfavorable to those published not long ago but having potential academic influence, and in order to eliminate this ＂unfairness＂, this paper introduces the citation interval into the algo- rithm. The empirical studies in the LIS field shows that the improved algorithm optimizes the evaluation results effectively, which is more favorable to finding newly-published literatures with potential academic influence compared with the classical PageRankrithm.%通过引文网络的结构特征，有效地识别科学文献的价值并建立某种序关系，为科学评价提供了有价值的参考，也丰富了科学评价的方法。针对传统PageRank算法在引文网络中得到的结果倾向于发表时间久的文献，而不利于发表时间较短但具有学术影响潜力的文献，为了消除这种“不公平”性，将引文间隔时间引入算法中。图书馆与情报学领域的实证研究说明改进算法有效地优化了评价的结果，相对于传统PageRank算法更有利于发现具有潜力的新发表的文献。