GraphTempo: An aggregation framework for evolving graphs

Abstract

Graphs offer a generic abstraction for modeling entities and the interactions and relationships between them. Since most real-world graphs evolve over time, there is a need for models to explore the evolution of graphs over time. We introduce the GraphTempo model that allows aggregation both at the attribute and at the time dimension. We also propose an exploration strategy for navigating through the evolution of the graph based on identifying time intervals of significant growth, shrinkage or stability. This exploration strategy would be useful for example for identifying time periods of multiple collaborations between specific groups in a cooperation network, or of declining contacts between specific groups in a disease propagation network. We evaluate the performance and effectiveness of our strategy using two real graphs.