Network science has emerged as one of the increasingly important research areas. Both generic and bipartite network structures are present in many real-world data in biological science, business, engineering, medical field, etc. In order to understand how network structures emerge and evolve over time, it is critical to carry out efficient estimation and make inferences of network parameters. The estimation and inference work can be accomplished by model-based clustering using exponential-family random graph models (ERGMs). In addition to ERGMs, we incorporate stochastic block models (SBMs) to estimate cluster membership for both static and dynamic bipartite networks. We further develop a variational expectation-maximization (VEM) algorithm to approximate the maximum likelihood of a bipartite network. The thesis will have an in-depth discussion of network analysis and its sub-fields, model-based clustering methods, computational algorithms, and numeric simulation.
