Exploiting Structural and Temporal Evolution in Dynamic Link Prediction

Link prediction in dynamic networks is an important task with many real-life applications in different domains, such as social networks, cyber-physical systems, and bioinformatics. There are two key processes in dynamic networks: network structural evolution and network temporal evolution, where the former represents interdependency between entities and their neighbors in the network at each timestamp, while the latter captures the evolving behavior of the entire network from the current timestamp to the next. Structural evolution generally assumes that a node is more likely to co-evolve with its neighbors in the near future. Temporal evolution focuses on the trend of network evolution as a whole, based on the accumulation of historical data. It is thus essential to use characteristics of both structural and temporal evolutions to emulate complex behaviors of dynamic networks. However, very few existing work considered both processes. In addition, real-life networks are often very sparse with limited observed links. A missing link between two nodes does not always imply that the two nodes do not have a relation in reality, especially when they share many common neighbors. Most existing methods only focus on the first-order proximity of networks, which is usually insufficient to capture the relationships among nodes.

In this work, we propose a novel framework named STEP, to simultaneously integrate both structural and temporal information in link prediction in dynamic networks. STEP first constructs a sequence of higher-order proximity matrices to better capture the implicit relationships among nodes. A regularized optimization problem is then formulated to model those higher-order proximity matrices along with additional structural and temporal constraints. Given the large scale of modern networks, we also develop an efficient block coordinate gradient descent approach to solve the optimization problem efficiently. STEP can be used to solve the link prediction problem in directed or undirected, weighted or unweighted dynamic networks. Extensive experiments on several real world datasets demonstrate that STEP can effectively model link propagation over entire time-varying networks and its superiority over some state-of-the-art algorithms.