| 基于贝叶斯网络的缺失数据处理 |
| |
| 引用本文： | 宫义山,董晨. 基于贝叶斯网络的缺失数据处理[J]. 沈阳工业大学学报, 2010, 32(1): 79-83 |
| |
| 作者姓名： | 宫义山 董晨 |
| |
| 作者单位： | 沈阳工业大学 信息科学与工程学院， 沈阳 110870 |
| |
| 基金项目： | 辽宁省教育厅科技计划项目(2008513) |
| |
| 摘 要： | 针对贝叶斯网络的数据缺失问题，提出了一种数据修补方法，该方法通过使用一个评分函数来对节点各个状态分别进行评分，得出各个状态对于所属的贝叶斯网络的得分，即各个状态与所属贝叶斯网络的匹配度，最终确定要补充数据的状态.实验证明，此方法可以在贝叶斯网络结构学习或者推理之前对数据进行修补，使其成为完整的数据，有效提高缺失数据的利用率.
|
| 关 键 词： | 贝叶斯网络 数据修补 评分函数 缺失数据 概率推断 数据处理 不确定信息理论 联合熵 |
Data patching method based on Bayesian network |
| |
| GONG Yi-shan,DONG Chen. Data patching method based on Bayesian network[J]. Journal of Shenyang University of Technology, 2010, 32(1): 79-83 |
| |
| Authors: | GONG Yi-shan DONG Chen |
| |
| Affiliation: | School of Information Science and Engineering, Shenyang University of Technology, Shenyang 110870, China |
| |
| Abstract: | A data pathcing method was proposed for the data missing problem of Baysian network. This method used a scoring function to score all nodes and obtained the scores of all states belong to Baysian network,where the scores are the matching degree of all states to Batsian network. The states of supplement data were determined. The experimental results show that this method can finish data patching before Baysian network strcuture learing or inferring,and thus the complete data can be obtained. Furtheomore,the ... |
| |
| Keywords: | Bayesian network data patching scoring function missed data probability inference data processing uncertain information theory joint entropy |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《沈阳工业大学学报》浏览原始摘要信息 |
|
点击此处可从《沈阳工业大学学报》下载全文 |
