| 1 1/2片联合树算法在动态贝叶斯网精确推理中的应用 |
| |
| 引用本文： | 周本达,王浩,姚宏亮. 1 1/2片联合树算法在动态贝叶斯网精确推理中的应用[J]. 计算机工程与应用, 2005, 41(14) |
| |
| 作者姓名： | 周本达 王浩 姚宏亮 |
| |
| 作者单位： | 合肥工业大学计算机科学技术系,合肥,230001;皖西学院数学系,安徽,六安,237012;合肥工业大学计算机科学技术系,合肥,230001 |
| |
| 基金项目： | 安徽省自然科学基金项目(编号:03042305)资助 |
| |
| 摘 要： | 基于动态贝叶斯网络处理动态不确定性问题的过程中推理是非常重要的,而推理算法的优劣决定着推理的执行效率。该文提出一种较简单的112片联合树算法,在不需要限制消去顺序且只作一次扩展的条件下构造联合树,所以算法简单且具有较小的复杂度。
|
| 关 键 词： | 动态贝叶斯网络(DBNs) 联合树 马尔可夫模型 |
An Application of 1 1/2 Slice Junction Tree Algorithm to the Exact Inference in DBNs |
| |
| Zhou Benda,Wang Hao,Yao Hongliang. An Application of 1 1/2 Slice Junction Tree Algorithm to the Exact Inference in DBNs[J]. Computer Engineering and Applications, 2005, 41(14) |
| |
| Authors: | Zhou Benda Wang Hao Yao Hongliang |
| |
| Affiliation: | Zhou Benda1,2 Wang Hao1 Yao Hongliang11 |
| |
| Abstract: | To solve the dynamic uncertainty problem which is based on DBNs,the efficiency of process is decided by inference algorithm.The paper presents a new 1 slices Junction Tree algorithm which don't need elimination order limited and constitute the Junction Tree only one times.So it's more simple and has a lower complexity. |
| |
| Keywords: | Dynamic Bayesian Networks(DBNs) junction tree Markov models |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
