基于Rough下和上近似的距离函数的数据约简  

A Model of Data Mining Based on Distance Function of Rough Lower and Upper Approximations

作　　者：刘群[1] 

机构地区：[1]武汉大学数学与计算机科学学院,湖北武汉430072

出　　处：《计算机与现代化》2000年第2期1-4,77,共5页Computer and Modernization

摘　　要：基于 Rough下和上近似集 ,本文定义了一个关于下和上近似的距离函数 ,并利用它可将属性值区间化 ,从而可以构造一个基于这个距离函数的数据挖掘模型。它将被应用于各类数据库中的数据约简。本文还讨论了带 Rough决策因子的决策规则 ,它是作决策的依据。它与MYCIN系统中的可信度因子类似 ,不同的是这个 Rough决策因子可通过数学公式计算出来 ,而可信度因子则不然。我们用这个模型开发了一个中医诊疗系统 ,目前正在临床和教学中试用。Based on rough lower and upper approximate sets , this paper defines the distance function with respect to lower and upper approximations, and the values of attributes are transformed into the interval values by the distance function. Whereby, we construct a model of data mining. The model will be applied to data reduction in various data base. It discusses the decision rules with rough decision fact in the paper, it is the evidence to take decision. It is similar to believable degree fact in MYCIN, different parts are that rough decision fact in the paper may be computed, but believable degree fact in MYCIN is not. An expert systems of diagnosis and treatment in Chinese traditional medicine are developed with the model, it is trying to be applied to clinic and teaching.

关 键 词：距离函数 数据函数 数据约简 中医诊疗系统 

分 类 号：R241[医药卫生—中医诊断学] TP311.13[医药卫生—中医临床基础]