Opinion leader detection: A methodological review

Abstract A social network as an essential communication platform facilitates the interactions of online users. Based on the interactions, users can influence or be affected by the opinions of others. The users being able to influence and shape the opinions of others are considered as opinion leaders. The problem of identifying opinion leaders is an important task due to its wide applications in reality, including product adoption for marketing and societal analytics. The problem has been attracting proliferating studies over the recent years. To overview and provide insights of the methodologies and enlighten the future study, we review the well-known techniques for opinion leader detection problems. These techniques are classified into descriptive approaches, statistical and stochastic methods, diffusion process based approaches, topological based methods, data mining and learning methods, and approaches based on hybrid content mining. The advantages and drawbacks of each method are systematically analyzed and compared, to provide deep understanding into the existing research challenges and the direction of future trends. The findings of this review would be useful for those researchers are interested in identifying opinion leaders and influencers in social networks and related fields.