基于粗糙集和决策树规则提取方法的研究毕业论文.doc
约44页DOC格式手机打开展开
基于粗糙集和决策树规则提取方法的研究毕业论文,摘要在许多行业的故障诊断过程中,针对故障诊断信息不一致的情况,通常采用一种基于粗糙集和决策树规则的提取方法,来提取简单有效的诊断规则。基于粗糙集和决策树规则的提取方法从定义的故障诊断决策系统出发,将故障诊断问题用一个具有不同简化层次的决策网络表示。在引入决策诊断规则的覆盖度概念后,推导出每个网络节点的诊断决策规则集。基...
内容介绍
此文档由会员 白痴学东西 发布
摘 要
在许多行业的故障诊断过程中,针对故障诊断信息不一致的情况,通常采用一种基于粗糙集和决策树规则的提取方法,来提取简单有效的诊断规则。基于粗糙集和决策树规则的提取方法从定义的故障诊断决策系统出发,将故障诊断问题用一个具有不同简化层次的决策网络表示。在引入决策诊断规则的覆盖度概念后,推导出每个网络节点的诊断决策规则集。基于粗糙集和决策树规则的提取方法具有良好的定性分析能力,因此,如何将粗糙集与决策树很好的融合,从而提取出有效的规则,已经是一个非常有研究性的问题了。
本文以粗糙集和决策树为研究多规则提取方法的基础。首先,介绍了论文的研究背景,国内外研究现状以及研究的意义。其次,对粗糙集理论进行详细研究,包括粗糙集中的一些基本的概念和粗糙集的简化。再次,将粗糙集理论应用到决策树中,应用准确度和覆盖度两个因素对提取的规则进行评价,最终提取简洁、高效的规则。最后,举出一个故障诊断方面的实例,将基于粗糙集和决策树规则的提取方法应用于实例中以解决问题,证明基于粗糙集和决策树规则的提取方法可对新的故障状态进行推理和决策,在故障诊断信息不完备的情况下也能得到相对满意的诊断结论。
关键词:粗糙集,决策树,覆盖度,故障诊断,决策规则
Abstract
In many industries the process of fault diagnosis, fault diagnostic information for inconsistencies, often using a decision tree based on rough sets and rule extraction method to extract the simple and effective diagnostic rules. Rules based on rough sets and decision tree extraction method from the definition of the starting fault diagnosis decision-making system, the fault diagnosis problem is simplified with a different level of decision-making network said. Diagnostic decision rules for the introduction of the coverage of concepts, each network node is derived diagnostic decision rules set. Rules based on rough sets and decision tree extraction method has good qualitative analysis, therefore, how to rough set and decision tree is a very good integration to extract the effective rules, has become one of the most researches of the problem.
This paper researches on rough set and decision tree based rule extraction method. First of all, introduced the research background, research status and research significance. Secondly, the detailed study of rough set theory, including some of the basic rough sets and rough sets concepts simplified. Again, the rough set theory applied to the decision tree, the application of accuracy and coverage of the two factors to eva luate the extracted rules, the final extraction of simple, efficient rules. Finally, to cite an instance of a fault diagnosis, the decision tree based on rough sets and rule extraction method is applied to solve the problem instances to prove that the rules based on rough sets and decision tree extraction method can be the fault of the new state of reasoning and decision-making in fault diagnosis case of incomplete information can be obtained relatively satisfactory diagnoses.
Key words: Rough sets, decision tree, coverage, fault diagnosis, decision rules
目 录
第1章 绪 论 1
1.1 论文研究的背景 1
1.1.1 粗糙集理论 1
1.1.2 决策树技术 2
1.2 国内外研究现状 2
1.3 研究的意义 3
1.4 本文的研究工作及结构组织 4
第2章 粗糙集理论的研究 5
2.1 引 言 5
2.2 粗糙集理论的基本概念 5
2.2.1 知识和知识库 5
2.2.2 信息系统和决策表 6
2.2.3 不可分辨关系 8
2.2.4 上、下近似集 9
2.2.5 依赖度和重要度 12
2.2.6 分明矩阵 13
第3章 基于粗糙集和决策树规则提取方法的研究 15
3.1 粗糙集的简化 15
3.1.1 数据离散化 15
3.1.2 属性简约与核 18
3.1.3 决策规则 20
3.2 决策树技术简介 21
3.3 基于粗糙集的决策树构造算法 23
3.3.1 普通的决策树算法简介 23
3.3.2 优化的决策树构造算法 26
3.4 决策规则提取及筛选 28
3.4.1 覆盖度 28
3.4.2 决策规则的提取及预测 30
3.4.3 故障诊断决策网络的推理 32
第4章 故障诊断实例研究 33
结 论 36
致 谢 37
参考文献 38
在许多行业的故障诊断过程中,针对故障诊断信息不一致的情况,通常采用一种基于粗糙集和决策树规则的提取方法,来提取简单有效的诊断规则。基于粗糙集和决策树规则的提取方法从定义的故障诊断决策系统出发,将故障诊断问题用一个具有不同简化层次的决策网络表示。在引入决策诊断规则的覆盖度概念后,推导出每个网络节点的诊断决策规则集。基于粗糙集和决策树规则的提取方法具有良好的定性分析能力,因此,如何将粗糙集与决策树很好的融合,从而提取出有效的规则,已经是一个非常有研究性的问题了。
本文以粗糙集和决策树为研究多规则提取方法的基础。首先,介绍了论文的研究背景,国内外研究现状以及研究的意义。其次,对粗糙集理论进行详细研究,包括粗糙集中的一些基本的概念和粗糙集的简化。再次,将粗糙集理论应用到决策树中,应用准确度和覆盖度两个因素对提取的规则进行评价,最终提取简洁、高效的规则。最后,举出一个故障诊断方面的实例,将基于粗糙集和决策树规则的提取方法应用于实例中以解决问题,证明基于粗糙集和决策树规则的提取方法可对新的故障状态进行推理和决策,在故障诊断信息不完备的情况下也能得到相对满意的诊断结论。
关键词:粗糙集,决策树,覆盖度,故障诊断,决策规则
Abstract
In many industries the process of fault diagnosis, fault diagnostic information for inconsistencies, often using a decision tree based on rough sets and rule extraction method to extract the simple and effective diagnostic rules. Rules based on rough sets and decision tree extraction method from the definition of the starting fault diagnosis decision-making system, the fault diagnosis problem is simplified with a different level of decision-making network said. Diagnostic decision rules for the introduction of the coverage of concepts, each network node is derived diagnostic decision rules set. Rules based on rough sets and decision tree extraction method has good qualitative analysis, therefore, how to rough set and decision tree is a very good integration to extract the effective rules, has become one of the most researches of the problem.
This paper researches on rough set and decision tree based rule extraction method. First of all, introduced the research background, research status and research significance. Secondly, the detailed study of rough set theory, including some of the basic rough sets and rough sets concepts simplified. Again, the rough set theory applied to the decision tree, the application of accuracy and coverage of the two factors to eva luate the extracted rules, the final extraction of simple, efficient rules. Finally, to cite an instance of a fault diagnosis, the decision tree based on rough sets and rule extraction method is applied to solve the problem instances to prove that the rules based on rough sets and decision tree extraction method can be the fault of the new state of reasoning and decision-making in fault diagnosis case of incomplete information can be obtained relatively satisfactory diagnoses.
Key words: Rough sets, decision tree, coverage, fault diagnosis, decision rules
目 录
第1章 绪 论 1
1.1 论文研究的背景 1
1.1.1 粗糙集理论 1
1.1.2 决策树技术 2
1.2 国内外研究现状 2
1.3 研究的意义 3
1.4 本文的研究工作及结构组织 4
第2章 粗糙集理论的研究 5
2.1 引 言 5
2.2 粗糙集理论的基本概念 5
2.2.1 知识和知识库 5
2.2.2 信息系统和决策表 6
2.2.3 不可分辨关系 8
2.2.4 上、下近似集 9
2.2.5 依赖度和重要度 12
2.2.6 分明矩阵 13
第3章 基于粗糙集和决策树规则提取方法的研究 15
3.1 粗糙集的简化 15
3.1.1 数据离散化 15
3.1.2 属性简约与核 18
3.1.3 决策规则 20
3.2 决策树技术简介 21
3.3 基于粗糙集的决策树构造算法 23
3.3.1 普通的决策树算法简介 23
3.3.2 优化的决策树构造算法 26
3.4 决策规则提取及筛选 28
3.4.1 覆盖度 28
3.4.2 决策规则的提取及预测 30
3.4.3 故障诊断决策网络的推理 32
第4章 故障诊断实例研究 33
结 论 36
致 谢 37
参考文献 38