Identifying effective influencers based on trust for electronic word-of-mouth marketing: A domain-aware approach

Because effective influencers in an online social network (OSN) can significantly affect consumers' purchasing decisions via trust among users in electronic word-of-mouth (eWOM) marketing, identifying these influencers with respect to user trust relationships has become increasingly important. However, many existing studies overlook the domain attribute of trust and the time-varying nature of social networks and only analyze a static snapshot of a user trust network (UTN). To address these issues and investigate this topic in the e-commerce context, this study proposes a research framework that takes into account the dimensions of trust, domain, and time. A time-varying hypergraph is developed to model the OSN using the time-varying features of multi-type relationships, and an algorithm is developed to extract a domain-aware UTN based on the time-varying hypergraph and user trust relationships. Reinforced by the dimensions of trust, domain, and time, a novel product review domain-aware (PRDA) approach is conceived that identifies effective influencers and categorizes them into three types, i.e., emerging influencers, holding influencers, and vanishing influencers, based on their popularity status across the life cycle. The experimental results from the Epinions dataset show that the PRDA approach outperforms both the social network-based influence-evaluating approach and the "popular author" approach.