Towards Detecting Influential Users in Social Networks

One of online social networks' best marketing strategies is viral advertisement. The influence of users on their friends can increase or decrease sales, so businesses are interested in finding influential people and encouraging them to create positive influence. Models and techniques have been proposed to facilitate finding influential people, however most fail to address common online social network problems such as fake friends, spammers and inactive users. We propose a method that uses interaction between social network users to detect the most influential among them. We calculate the relationship strength and influence by capturing the frequency of interactions between users. We tested our model in a simulated social network of 150 users. Results show that our model succeeds in excluding spammers and inactive users from the calculation and in handling fake friendships.