基于粒计算的二进制矩阵及在决策树算法的应用  

Binary System Matrix Based on Granular Computing and Its Applications in Decision-tree

作　　者：徐健锋[1,2] 刘斓[2] 邱桃荣[2] 刘清[2] 

机构地区：[1]南昌大学软件学院,江西南昌330047 [2]南昌大学信息工程学院,江西南昌330003

出　　处：《广西师范大学学报（自然科学版）》2008年第3期157-160,共4页Journal of Guangxi Normal University:Natural Science Edition

基　　金：国家自然科学基金资助项目(60173054);江西省教育厅科技项目(GJJ080382008)

摘　　要：介绍了基于粒计算的二进制信息粒矩阵,给出了粒计算粒度的概念、粒的关联度运算、关联粒度矩阵和条件属性重要度等几个基于二进制信息粒计算的基本定义,并基于上述定义提出一种新的基于粒计算的决策树分支和剪枝算法在气象预报中的应用,实例结果证明了所提出的算法是有效和可行的。A binary system in information granules matrix based on granular computing is induced to illustrate several definitions of binary system in information granules,such as ,the concept of granularity, relevance computing of granule, association granularity matrix, importance degree of conditional attribute,etc. Then a new Binary numeral matrix decision tree algorithm based on granular computing is presented and is applied to weather forecast ,this method proves to be practical and feasible.

关 键 词：粒计算 信息粒 二进制矩阵 决策树 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]