Integration of Social Network Analysis and Text Mining for Identification of Influential Opinion Leaders

Opinion leaders are often able to influence others in a specific domain by providing new information, ideas and opinions. Identification of opinion leaders has become one of the most important tasks in the field of word of mouth mining. Most existing research tasks for the identification of opinion leaders are based on quantitative approaches, such as social network analysis, clustering, and statistical methods. However, the content of word of mouth documents is also significant, as opinion leaders often post informative and useful documents with the attributes of novelty, expertise and richness of information. The quality of documents also indicates the possibility of the poster being an opinion leader. Thus, this research proposes two integrated strategies of quantitative and qualitative techniques to evaluate features of opinion leaders. Based on a real-world online discussion forum, the involvement approach is the best of the quantitative approaches. The text mining-based approach and the first integrated approach have the same performance as the involvement approach. The performance of the second proposed integrated approach in identifying opinion leaders achieves the best performance. JEL classification numbers: M15, O35.