摘要知识约简是粗糙集理论的核心内容，粗糙集理论(Rough Set)作为智能信息处理技术的一个新成果，是由波兰科学家Pawlak教授所提出的对不完整数据进行分析、推理、学习、发现的新方法[2,3]。运用知识约简的方法对城市轨道交通案例进行约简。23124
本文首先研究了粗糙集的基础理论知识，讨论了粗糙集的定义及粗糙集中有关知识分类、知识约简等内容；其次探讨了决策表约简中的四种属性约简算法和四种属性值约简算法。现今，大部分决策问题都可以通过决策表形式地表达出来，因此它在工程应用中越来越重要。
在上述理论研究的基础上，本文以城市轨道交通地铁为背景，分别通过熵值法、K-Fold交叉验证法和K-NN法进行了知识约简的实际应用。本文通过对地铁属性详细分析，构建了地铁属性的知识约简模型，提出了地铁系统知识的获取与表达的方法。最后三种用于地铁参数属性约简的方法进行了对比分析。全文旨在说明知识约简在决策简化过程中的实际应用意义。
关键词  城市轨道交通  知识约简  熵  粗糙集
毕业论文外文摘要
Title A case study of knowledge reduction method of city rail traffic energy consumption
Abstract
Knowledge reduction is the core of Rough Set. Rough Set as a new outcome of intelligent information processing technology, and it’s a new method of analyzing, reasoning, learning and discovering the incomplete data proposed by Pawlak who is a Polish scientist Professor[2,3].Applying knowledge reduction to carry out the reduction to the city rail transit.
This paper studies the basic knowledge of Rough Set, discusses the definition ,knowledge classification, and knowledge reduction of Rough Set; Secondly the paper discusses four algorithms on attribute reduction and four algorithm values on attribute reduction of the decision table reduction. Nowadays, most problems of decisions can be expressed in the form of decision tables , so it is increasingly important in engineering applications .
On the basis of these theoretical studies, this paper applies the entropy method, K-Fold cross-validation method and K-NN method on the practical application of knowledge reduction under the background of urban rail transit subway. This paper uses detailed analysis of the properties of the subway to construct the model of subway attributes knowledge reduction, and propose a method to obtain and express the knowledge of the subway system. Finally we compare and analyse the three Subway parameter attribute reduction. The full text is intended to illustrate the significance of the practical application on knowledge reduction in the process decision reduction.
Keywords  city track traffic, knowledge reduction, entropy, rough set
 目次
1 绪论    1
1.1 研究背景及意义    1
1.2 本文研究内容    2
2 粗糙集理论    4
2.1 集合间的等价关系    4
2.2 知识的分类观点    4
3 决策表的约简    6
3.1 知识的约简    6
3.1.1 一般性的约简    6
3.1.2 相对性的约简    7
3.1.3 知识的依赖性    7
3.2 信息系统及表示    8
3.3 属性的依赖性    9
3.4  属性约简的算法    10
3.4.1 一般约简算法    10
3.4.2 K-Fold交叉验证法    10
3.4.3 K-NN法    11
3.4.4 熵值法    12
3.4.5 基于属性重要性的启发式属性约简算法    12
3.4.6 基于互信息的属性约简算法——MIBARK算法    14
3.5 属性值的约简算法    15 城市轨道交通能耗案例知识约简方法研究:http://www.youerw.com/jisuanji/lunwen_16058.html