Modelling sparsity, heterogeneity, reciprocity and community structure in temporal interaction data
This work addresses the challenge of accurately modeling social interactions in network data, which is important for researchers in network science and machine learning, though it appears incremental as it builds on existing methods like compound random measures.
The authors tackled the problem of modeling temporal interaction networks by proposing a novel class of Hawkes process models that incorporate sparsity, degree heterogeneity, community structure, and reciprocity, and demonstrated that their model outperforms competing approaches in link prediction tasks on real-world data.
We propose a novel class of network models for temporal dyadic interaction data. Our goal is to capture a number of important features often observed in social interactions: sparsity, degree heterogeneity, community structure and reciprocity. We propose a family of models based on self-exciting Hawkes point processes in which events depend on the history of the process. The key component is the conditional intensity function of the Hawkes Process, which captures the fact that interactions may arise as a response to past interactions (reciprocity), or due to shared interests between individuals (community structure). In order to capture the sparsity and degree heterogeneity, the base (non time dependent) part of the intensity function builds on compound random measures following Todeschini et al. (2016). We conduct experiments on a variety of real-world temporal interaction data and show that the proposed model outperforms many competing approaches for link prediction, and leads to interpretable parameters.