基于属性重要性约简算法的实现.doc

毕业设计
设计题目基于属性重要性约简算法的实现
学生姓名齐欣
学号 0105010220
专业班级计算机科学与技术 01-2班
指导教师胡学钢
2005年 6月 19 日
基于属性重要性约简算法的实现
Implement an Algorithm
For Reduction of Knowledge
Based On Significance of Attributes
齐欣
合肥工业大学学士学位论文
2005 . 6
摘要
。用粗糙集理论进行数据挖掘,抽取知识规则,最重要的一点就是基于粗糙集的属性约简。通过约简操作,降低属性维数,总结出适于决策支持的知识规则,是粗糙集理论的重要应用之一。约简,是通过删除知识库中多余的部分(等价关系)或多余基本属性集来保留知识库中的重要属性,从而删除知识库中不必要的知识,只保留那些真正有用的部分。目前求解知识约简的算法主要有两种:一种是利用辨识矩阵构造区分函数,。该算法从最重要的属性开始取,直到取出的属性构成一个约简为止,从而保证了求得的属性约简一定是最小约简。
关键词:粗糙集理论,分辨矩阵,属性重要性,属性约简
Abstract
Rough set theory, initialized by Polish scientist in early 1980’s, has been proved to be an excellent mathematical tool dealing with uncertain and vague description of objects, whose basic idea is to derive classification rules of conception by knowledge. Carries on the data mining with the rough set theory, extracts the knowledge rules, most important one is attribute reduction based on the rough set. By means of reductional operation, reduces the attribute dimensionality, summarize of the knowledge rule which is suitable for the policy-making support, is one of important applications of the rough set theory. Reduction is a kind of operation that by means of deleting the otiose part in repository (equivalence relation) or reduct basic attribute set in order to preserving important attributions those in repository, consequently to delete the needless knowledge which in repository, and nothing but useful attribute could be saved. Presently, there are two kinds of arithmetic for reduction of knowledge. One is by using discrimination matrix construct discrimination function. The other is an algorithm based on significance of attributes. The main content of this thesis is program to actualize a kind of algorithm to reduction of knowledge based on significance of attributes which start fetch attribution from foremost one, and not stop fetching until the attribution