Topical Authority-Sensitive Influence Maximization.

Influence maximization has been widely studied in social network analysis. However, most existing works focus on user’s global influence while ignoring the fact that user’s influence varies with different topics. Even though a few works take topics into consideration, they all neglect the authority of users on a specific topic which is also a very important indicator when selecting seed nodes. In this paper, we propose a new Topical Authority-sensitive Independent Cascade model (TAIC) by introducing user’s authority on a given topic, and based on which, a Topical Authority-sensitive Greedy algorithm (TAG) is presented to try to find the most influential nodes on the given topic. We also propose a new metric, Influence Spread of seed set on a Given Topic (ISGT) to evaluate the performance of our proposed model and algorithm. Experiments on two real-world datasets show that our proposed TAG algorithm performs continuously better in finding nodes with higher influence on a given topic and has better time efficiency than other baseline algorithms.