Node Aggregation of Bayesian Network Based on Chain Graph
Weihua Li · Jisuanji gongcheng · 2004
Based on chain graph, we present an algorithm to aggregate the nodes in a Bayesian network. Firstly, Bayesian network was transformed into chain graph. The relation between nodes in chain graphs partitioned nodes into some subset of nodes viewed as domains. Then, each domain was aggregated into a new node in the new Bayesian network. Finally, By refining the parameters, we constructed the Bayesian network again. This procedure make the probabilistic relation between domains more clear, and at the same time, make the structure of the Bayesian network can be optimized.