摘要
研究算法改进,提高计算性能,贝叶斯网络是解决不确定性问题的一种有效方法,在很多领域得到了广泛应用。参数学习是贝叶斯网络构建的重要环节,但含隐变量、连续变量的参数学习是非常困难的。为解决上述问题,提出了一种人工鱼群算法的贝叶斯网络参数学习方法,并进一步通过调整人工鱼随机移动速度的方法提高了算法的收敛性能和速度。最后,将参数学习方法在由Noisy-Or和Noisy-And节点组成的贝叶斯网络中进行了仿真,仿真结果表明了参数学习方法,特别是改进后方法的可行性和优越性。
Bayesian network is an effective method to solve uncertainty problem, and has been widely used. Parameter learning is an important step for building a Bayesian network, but it is difficult for the network with hidden variables or continuous variables. To solve this problem, an artificial fish swarm algorithm based method was presented for learning Bayesian network parameters. Furthermore, the algorithm＇ s convergence and speed were improved by adjusting the speed of random movement. Finally, this method was simulated in the network composed of Noisy-Or and Noisy-And nodes. The result shows the feasibility and superiority of the improved method.
出处
《计算机仿真》
CSCD
北大核心
2012年第1期184-187,共4页
Computer Simulation
基金
华北电力大学青年教师科研基金(200911021)
关键词
贝叶斯网络
人工鱼群算法
参数学习
Bayesian network
Artificial fish swarm algorithm
Parameter learning
