Topology-based Algorithm for Users' Influence on Specific Topics in Micro-blog

摘要

In recent years, users' influence in the Micro-blog has been studied extensively in the field of social network. Most previous studies are based on a single factor including followers, followees and tweets. However, Micro-blog is different from other social network services. It is difficult to truly reflect the scale and nature of the users' influence with a single factor in the service. In this paper, we focus on the ranking strategies of users for specific topics in terms of two major factors (users' activities and quality of tweets), which are highly relevant to the social network properties of the Micro-blog users and the properties of the Micro-blog itself. As a step forward, we propose a Weighted Ranking Algorithm (WRA) through analysing Kendall's γ correlation of different rank lists by the two factors. The experimental results based on a large-scaled real dataset from Sina Micro-blog show that the proposed WRA algorithm is more effective than the existing algorithms.