Link Predictions for Social Networks in an Online Health Community

Online social networks are ubiquitous in our daily life and offer different ways for people to interact with each other. Link prediction aims at predicting future social connections or interactions between two users within a social network. This project implements different link prediction algorithms and evaluates their performance for online social networks among users of an online health community for smoking cessations. The social networks were based on users’ online interactions via four communication channels: blog comments, message boards, group discussions and private messages. The outcome of this study will help to provide insights into the design of recommender systems in such online social networks, and to improve user experience and engagement in online health communities.