| 基于贝叶斯网络的过度训练诊断 |
| |
| 引用本文： | 马振龙.基于贝叶斯网络的过度训练诊断[J].农业网络信息,2007(12):56-59. |
| |
| 作者姓名： | 马振龙 |
| |
| 作者单位： | 北京联合大学,北京,100101 |
| |
| 摘 要： | 针对贝叶斯网络在具体应用时需根据实际应用问题来建立相应的贝叶斯网络模型的问题,采用Java语言设计并开发了一套贝叶斯网络预测诊断系统,用来实现因果推理和诊断推理.重点讨论了系统的网络数字化、网络学习、网络推理等关键问题.利用该平台建立了过度训练的诊断网络,通过试验证明得出,此套系统简单易用、通用性强,可应用于不同的研究领域.
|
| 关 键 词： | 贝叶斯网络 预测 诊断 过度训练 贝叶斯网络 过度训练 诊断推理 bayesian networks based 研究 通用性 验证 平台 利用 网络推理 网络学习 网络数字化 诊断系统 重点 因果推理 网络预测 开发 语言设计 Java |
| 文章编号： | 1672-6251(2007)12-0056-04 |
| 修稿时间： | 2007年10月29 |
Diagnose of overtraining based on bayesian networks |
| |
| MA Zhen-long.Diagnose of overtraining based on bayesian networks[J].Agriculture Network Information,2007(12):56-59. |
| |
| Authors: | MA Zhen-long |
| |
| Abstract: | In order to solved the question that Bayesian Networks must be constructed for a specific object in the practical application,the author developed a system for the use of causal and diagnostic inference.Some improtant problems for realization of the system were discussed in this paper,such as digitization,learning and inferenceof network.A Bayesian network for overtraining was constructed under this system.The practical application confirms that this system provided an effective and powerful function so that it can be applied to different fields. |
| |
| Keywords: | Bayesian networks Predict Diagnose Overtraining |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
