Structure Learning Of BN Based On Improved AFS And ABC Algorithm

Bayesian network is a graphical network, and it is the combination of graph theory and probability theory. It is used to express probability distribution between the node and reasoning. In the past dozens of years, the Bayesian network model had achieved good effects in uncertain systems applications, and it is widely used in the research of uncertainty. Bayesian network research mainly focused on the learning of Bayesian networks, especially the structure learning of Bayesian network. But now the research shows that the Bayesian network structure learning is a NP problem, so exploring the effective learning methods for structure learning is the core of the Bayesian network theory research.In this paper, the Bayesian network structure learning has been conducted in-depth research. The artificial fish swarm algorithm (AFSA) and artificial bee colony algorithm (ABCA) was improved and introduced into the study of Bayesian network structure to form a new learning method. This article carries on the detailed introduction to these researches, and the following is a brief description of this work.First of all, the paper expounds the basic theory of Bayesian network, and the research status. Then, this paper introduces the learning of Bayesian networks, particularly with regard to the development of structure learning, which leads to the research background of this article.Second, detailed introduces the AFSA used in Bayesian network structure learning. Here it should be pointed out that the improvement for AFSA, including cloud-based adaptive theory, genetic algorithm, and then introduce the method of learning Bayesian network structure based on hybrid genetic and fish swarm algorithm.Then, the differential evolution algorithm is introduced into the framework of the ABC algorithm, combined with the previous proposed Cloud-based adaptive theory, and then proposes the hybrid differential evolution and artificial bee colony algorithm.Finally, the paper summarized the study of this article and proposed my personal view about the study of the structure learning of Bayesian network, and then discussed the structure learning research of Bayesian network and its development in the future.