Predicting Opinion Leaders in Word-of-Mouth Communities

Opinion leaders play a major role in shaping potential customers’ minds in word-of-mouth communities. Prior research on opinion leaders in word-of-mouth communities has focused mostly on structural characteristics of social networks of the leaders, ignoring the predictive potential of the actual text that they write. We argue that textual characteristics of reviews, along with reviewer characteristics, could be used to predict opinion leadership. In this paper, we propose a predictive model to identify opinion leaders in online word-of-mouth communities using both review and reviewer characteristics. The results from our study indicate that the predictive performance of the classification models built using the proposed predictors is much better than that of a baseline bag-of-words model built using the actual review text.