计算机应用与软件Issue(11)：178-182,5.DOI:10.3969/j.issn.1000-386x.2014.11.044
基于粒子群优化建模的贝叶斯网络结构学习方法
A BAYESIAN NETWORKS STRUCTURE LEARNING METHOD BASED ON PARTICLE SWARM OPTIMISATION MODELLING
李东灵1
作者信息
- 1. 商丘职业技术学院 河南 商丘 476000
- 折叠
摘要
Abstract
Bayesian networks ( BNs ) have good performances when used for representing information and reasoning under uncertain circumstances, but due to the complexity in its structure search space, a structure learning BN from a dataset is in general regarded as an NP-hard problem.Based on this, we propose a novel BN structure learning approach which is based on particle swarm optimisation ( PSO) modelling.In order to learn a structure of Bayesian network, the approach first uses PSO to search in sorting space, and then calculates the fitness of each ordering by running the K2 algorithm.Every ordering will have a corresponding network structure, and the approach will feed back the score of this network.Simulation results show that the approach can produce better network stabilities on the networks with different types compared with other BN structure learning algorithms in databases with different scales.关键词
贝叶斯网络/粒子群优化/K2算法/结构学习Key words
Bayesian networks/Particle swarm optimisation/K2 algorithm/Structure learning分类
信息技术与安全科学引用本文复制引用
李东灵..基于粒子群优化建模的贝叶斯网络结构学习方法[J].计算机应用与软件,2014,(11):178-182,5.
