A Bayesian Network Structure Learning Method Based on Ant Colony Algorithm
Fengshan Wang, Wanhong Zhu · 2010
To find the implied dependency relationships and knowledge representation from sample data, a Bayesian Network structure learning method was proposed on the basis of ant colony algorithm, which provided support for the modeling of complex decision-making tasks. Algorithm design was presented after the formal description of Bayesian network structure learning problems. Accordingly, a Bayesian network structure learning rules were built, including node state transferring rule, node sequence scoring rule, pheromone inspired strength calculating rule, pheromone updating rule, process controlling rule, network structure establishing rule, etc. Finally, the important value of algorithm applied in complex decision-making fields was effectively verified with the example of construction about damage assessing system for a certain military engineering.