具有时间反馈的PageRank改进算法

An improved PageRank algorithm with time feedbacking

针对某一类网页(比如新闻网页)在互联网上发布时间越长,其信息的重要性将随之下降这一事实,在传统的PageRank算法中加入时间反馈因子,实现网页因发布时间的长短,其PageRank值也随之上下浮动.并采用Seidel迭代算法加速迭代收敛过程.实验结果表明,改进后的算法在计算这类与发布时间相关的网页的PageRank值时,符合人们的一般期望,是有效的.Seidel迭代算法有利于提高算法效率.

<Abstrcat> PageRank is a web page ranking algorithm proposed by Google, a well known search engine. The algorithm is an iterative process that determines web page ranking based on page link structure, or co-citation. PageRank is a successful, but not a perfect algorithm. For instance, an older page is always an important page because the more older it is, the more link-in pages it has. So a new page is usually not important. For this, we first integrated page time information with PageRank calculation, and then employed Seidel's method to speed up the convergence of the iteration process. Experimental results show that the new algorithm is good and reasonable.