The analysis of user engagement dynamics in online question and answer communities
Kulkarni, Koustubh S.
MetadataShow full item record
Internet-based Question and Answer (Q&A) forums are effective in assembling and providing online users with access to large databases of subject expertise. Recently, such forums have enjoyed widespread growth. In an industrial setting, leveraging the information collection capability of social Q&A forums for product development or marketing could result in greater profits and growth for companies. The key question in such efforts is how to choose and implement a targeted information collection strategy under a given limited budget. To this end, it is important to understand the mechanisms of user engagement into forum thread-growing activities. This thesis presents a framework to quantify user engagement as the probability that a user will contribute to a thread under given circumstances, expressed via the features capturing the fit of the thread-related topics to the user's interests, engagement of the user's friends, and global popularity of the thread. The conducted computational experiments based on synthetic data test the proposed model's ability to fit the data, and reveal which traits of the real-world engagement patterns are well-captured by the model.