| 基于信息论和免疫遗传算法学习贝叶斯网络结构 |
| |
| 引用本文： | 张薇薇,卢玉贞,孙建英.基于信息论和免疫遗传算法学习贝叶斯网络结构[J].大连海事大学学报(自然科学版),2006,32(3):115-118. |
| |
| 作者姓名： | 张薇薇 卢玉贞 孙建英 |
| |
| 作者单位： | 大连海事大学数学系 辽宁大连116026 |
| |
| 摘 要： | 提出一种新的实现贝叶斯网络(BN)结构学习的方法,即由信息论和免疫遗传算法相结合构造最优贝叶斯网络结构.首先,通过信息论迅速搜索较为可能的网络空间,构造一个候选网络,然后用免疫遗传算法进行网络精简,得到一个最优的网络结构.不仅可降低计算复杂度,提高贝叶斯网络结构学习的效率,而且避免了传统启发式算法易陷入局部最优解的问题.实例计算证明了其可行性.
|
| 关 键 词： | 贝叶斯网络 交互信息 免疫遗传算法 信息论 |
| 文章编号： | 1006-7736(2006)03-0115-04 |
| 收稿时间： | 2006-03-30 |
| 修稿时间： | 2006年3月30日 |
Bayesian network based on information theory and immune genetic algorithm |
| |
| ZHANG Wei-wei,LU Yu-zhen,SUN Jian-ying.Bayesian network based on information theory and immune genetic algorithm[J].Journal of Dalian Maritime University,2006,32(3):115-118. |
| |
| Authors: | ZHANG Wei-wei LU Yu-zhen SUN Jian-ying |
| |
| Institution: | Math. Dept. , Dalian Maritime Univ. , Dalian 116026, China |
| |
| Abstract: | This paper proposed a new method to study Bayesian network(BN),which is based on the methods of information theory and immune genetic algorithm to find out the optimal BN structure.First it used the information entropy theory to search the possible network structure quickly.Then,found out the best BN structure using the immune genetic algorithm.In this way,it not only reduced the computing complexity and increased the effectiveness of studying BN structure,but also avoided the problems of local convergence in the traditional methods.The experiment reveals the effectiveness of the new method. |
| |
| Keywords: | Bayesian network mutual information immune genetic algorithm information theory |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
