基于改进PageRank算法的微博影响力模型研究

摘要

As PageRank algorithm is mainly based on the relationship between the page links to calculate the importance score of the page, it cannot be directly applied to the calculation of the user's influence on microblogging network. Based on PageRank algorithm, the indicators such as active degree of user, quality of the blog and structural similarity of location and so on are introduced to establish qualified users influence model: Influence_Index, which is suitable for the microblogging network users. It was found that four of the top fifteen users obtained by the Infuluence_Index modelappears in the five most influential user's list, which is generally acknowledged by users of the website. This shows the model has high accuracy and can better identify the influential network users, it deserves to be promoted further. Besides, we also find that the increase of number of the original blog, attention, and fans has a significant role on enhancing the user's influence.%由于PageRank算法主要是根据网页之间的相互链接关系来计算网页的重要性得分，因此不能直接应用到微博网络来计算用户的影响力，从而本文在PageRank算法的基础上，引入微博网络中用户的活跃度、博文质量以及关于位置的结构相似性等指标，建立适用于微博网络的用户影响力量化模型———Influence_Index。通过实验发现Influence_Index模型结果的影响力排名前十五的用户，其中有4位出现在被该网络用户公认的5位最具影响力用户的名单中。由此看出该模型的结果具有较高的准确性可以较好的识别出网络中具有影响力的用户，具有较高的推广意义。同时发现增加原创博文量、关注数量和粉丝量对提升用户的影响力有着显著作用。