Skip to Main Content
Learning Bayesian networks can be examined as the combination of parameter learning and structure learning. Parameter learning is estimation of the conditional probabilities (dependencies) in the network. Structural learning is the estimation of the topology (links) of the network. The structure of the network can be known or unknown, and the variables can be expressed as complete and incomplete data. In this paper, we introduce two cases of learning Bayesian networks from complete data: known structure and unobservable variables and unknown structure and unobservable variables.