基于网页排名算法的敏感性分析

摘要

The sensitivity analysis of Web page ranking algorithm can achieve a further understanding on principle and condition scope of the score of welcome degree which the algorithm model given. Different changes in parameters will result in different degrees of sensitivity. Based on this problem, this article discusses the sensitivity of PageRank and HITS by mathematical content analysis of the algorithm. On the basis of analyzing the dependence of matrix G on the ratio parameterα, hyperlink matrix H and personalized vector νT, this article analyzes the influence of three specific parameters on PageRank vector. Finally, it considers the sensitivity of HITS.%对网页排名算法的敏感性分析，能够进一步了解关于算法模型所给出的欢迎度评分的原理和条件范围。基于参数的不同变化会导致不同程度的敏感性这一问题，本文通过对算法的数学内容分析，研究PageRank和HITS的敏感性问题。在分析矩阵G对于比例参数α、超链接矩阵H和个性化向量vT的依赖性的基础上，分析了3个特定参数对PageRank向量的影响，最后，对HITS的敏感性进行分析。