Research Publications

Online forums are rich sources of information about users' communication activity over time. Finding temporal patterns in communication records can advance our understanding of the dynamics of conversations. The main challenge of temporal analysis in this context is the complexity of forum data. There can be thousands of interacting users with dozens of possible user features. It is not obvious which features to choose, and it is not obvious how to dene features that evolve over time.We propose to decouple temporal information about users into sequences of users' events and inter- event times.We develop a new feature space where we represent the sequences of events as paths, and we model the distribution of the inter-event times. We study 5; 675 users from three Internet forums, and discover novel patterns in user communication. We nd that users tend to exhibit consistency over time. Further, in our feature space, we observe areas that represent unlikely types of users' temporal behavior. We then derive a feature representation to describe forums. Using our forum features we discover an alternative clustering for one of the previously studied datasets.