基于概率粗糙集模型的属性约简方法研究

项目名称： 基于概率粗糙集模型的属性约简方法研究

项目编号： No.61502419

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 其他

项目作者： 马希骜

作者单位： 浙江工商大学

项目金额： 16万元

中文摘要： 与经典粗糙集模型中的属性约简相比，由于在概率粗糙集模型中引入了阈值，使得概率粗糙集属性约简过程中所用的度量不再满足单调性，而度量的非单调性导致概率粗糙集模型中的属性约简变得更加复杂。本项目针对概率粗糙集模型中度量的非单调性问题，深入研究度量的非单调性给概率粗糙集属性约简带来的本质影响，在此基础上通过研究满足单调性的度量构建科学的概率粗糙集属性约简方法。其主要研究内容包括：提出概率粗糙集的不确定性度量方法以及基于概率粗糙集不确定性度量的属性约简方法；分析不同概率粗糙集属性约简定义之间的关系，并通过满足单调性的度量方法构建科学的概率粗糙集属性约简原则；建立概率粗糙集属性核模型，提出概率粗糙集属性核的计算方法；研究完备有效的概率粗糙集属性约简算法。本项目的研究成果将为概率粗糙集属性约简理论提供全新的发展与洞察力，对进一步促进该理论的发展和实用化有着极其重要的理论和现实意义。

中文关键词： 属性约简；粗糙集模型；决策信息系统；概念近似；信息系统

英文摘要： Compared with attribute reduction in classical rough set model, the monotonicity of the measures for attribute reduction in probabilistic rough set model does not hold because of introducing the threshold values. The non-monotonicity of the measures makes the attribute reduction in probabilistic rough set model more complicated. To deal with the non-monotonicity problem of the measures in probabilistic rough set model, the essential effect of the non-monotonicity of the measures on attribute reduction in probabilistic rough set model will be deeply studied in the project. On this basis, a scientific method for attribute reduction in probabilistic rough set model will be established by studying the monotonic measures. Its key research points are: to propose methodologies dealing with uncertainty measure problem in probabilistic rough set model, and provide attribute reduction approaches based on uncertainty measures in probabilistic rough set model; to analyze the relationship between the different definitions of attribute reducts, and establish the scientific principle of attribute reduction by studying monotonic measures in probabilistic rough set model; to build the attribute core model and present the computing method of attribute core in probabilistic rough set model; to provide a systematic study of attribute reduction algorithm in probabilistic rough set model. The research results of this project will provide a new insight into the attribute reduction in probabilistic rough set model, and have the extremely important theory and the practical significance to promote the development of attribute reduction theory in probabilistic rough set model.

英文关键词： attribute reduction;rough set model;decision information system;concept approximation;information system