基于模糊概念层核属性集确定算法及仿真研究  

An Ascertainment Arithmetic of Core Attribute Set Based on Fuzzy Conceptual Hierarchy and Its Simulation Research

作　　者：李茹[1] 张丽芳[1] 褚诚缘[2] 邓嵩[3] 

机构地区：[1]山西大学计算机与信息技术学院,山西太原030006 [2]中国科学院计算技术研究所,北京100080 [3]山西财经大学,山西太原030006

出　　处：《计算机仿真》2006年第4期99-102,共4页Computer Simulation

基　　金：国家863高技术研究发展计划项目资助(编号:2004AA115460)

摘　　要：大型数据库中核属性集的确定是生成分类规则算法的重要步骤。该文在结合粗糙集理论和层次聚类方法的基础上,提出运用模糊概念层方法对原始数据进行模糊化处理,排除边缘数据的干扰作用并简化可区分矩阵的构造,从而得出核属性集。并以我国湖库在藻类繁殖高峰期实测浓度数据为例,快速、准确地得出湖库富营养化的主要指标。仿真试验结果表明,该算法简单,易于实现,在正确得出核属性集的基础上可以减少计算量,从而较为快速有效地得出推理规则。The ascertainment of the core attribute set of large database is an important process of generating the classifying rules. With the combination of rough set theory and hierarchical clustering method, this paper blurs the original data using fuzzy conceptual hierarchy method, eliminates the interference effect of marginal data, reduces the constitution of discernibility matrix and gets the core attribute set. Finally we got the main indices of eutrophication of lakes fleetly and exactly by analyzing the measured concentration of lakes in their algae breeding peak time. The simulation result indicates that the arithmetic is simple to be implemented and could increase computational efficiency and get the inference rules better and truly.

关 键 词：模糊概念层 核属性集 数据挖掘 知识发现 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]