| 基于结点排序的贝叶斯网络结构学习 |
| |
| 引用本文： | 王双成. 基于结点排序的贝叶斯网络结构学习[J]. 计算机工程与应用, 2005, 41(18): 11-12,187 |
| |
| 作者姓名： | 王双成 |
| |
| 作者单位： | 上海立信会计学院信息科学系,上海,201600 |
| |
| 基金项目： | 国家自然科学基金资助项目(编号:60275026)，吉林省自然科学基金项目(编号:20030517-1) |
| |
| 摘 要： | 给出了变量之间k阶分类能力的概念及计算方法，并证明了k阶分类能力就是k阶分类正确率，以及k阶分类能力和条件独立性的等价性，在此基础上构造出基于分类能力的贝叶斯网络结构打分函数，同时结合依赖分析方法和打分-搜索方法建立了有效的贝叶斯网络结构学习方法，实验结果显示该方法能够有效地进行贝叶斯网络结构学习，并使学习得到的结构倾向于简单化。
|
| 关 键 词： | 贝叶斯网络 结构学习 分类能力 打分函数 |
| 文章编号： | 1002-8331-(2005)18-0011-02 |
Learning Bayesian Network Structure Based on Nodes Ordering |
| |
| Wang Shuangcheng. Learning Bayesian Network Structure Based on Nodes Ordering[J]. Computer Engineering and Applications, 2005, 41(18): 11-12,187 |
| |
| Authors: | Wang Shuangcheng |
| |
| Abstract: | The concept and algorithm of k order classification ability is presented.It is proved that the k order classification ability is the k order classification accuracy and it is equivalent with conditional independency.The scoring function of Bayesian network structure is developed.And a new method of learning Bayesian network structure is set by combining dependency analysis and search & scoring.Experimental results show that this method can effectively learn Bayesian network structure from data and learned structure is inclined to simplification. |
| |
| Keywords: | Bayesian network structure learning classification ability scoring function |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
