基于贝叶斯网络的数据挖掘技术  

A Data-minning Technology Based on Bayesian Network

作　　者：陈秀琼[1] 

机构地区：[1]三明高等专科学校计算机科学系,福建三明365004

出　　处：《三明高等专科学校学报》2004年第2期47-52,共6页Journal of Sanming College

摘　　要：从海量数据中挖掘有用的信息为高层的决策支持和分析预测服务,已成为网络时代人们对信息系统提出的新的需求,但我们发现数据处理和数据的提炼技术是匮乏的。起源于贝叶斯统计学的贝叶斯网络以其独特的不确定性知识表达形式、丰富的概率表达能力、综合先验知识的增量学习方法等特性表示了客体的概率分布和因果联系,成为当前数据挖掘众多方法中最为引人注目的焦点之一。本文首先对贝叶斯网络、贝叶斯网络推理和贝叶斯网络学习进行综合性的阐述,然后讨论其在数据挖掘中的应用和优势。We found that it has been a new requirement for information system to mine useful information which is servered for making hight policy and forecast in internet ages. However, the tool to mining useful information is scarce. Bayesian network, which stemed from Bayesian statistics, could conveny object's probability and causality with its unique form to express uncertain knowledge and its strong ablitity to conveny probability and its synthetical prior knowledge to increasing learning ,and which has been a focus compared with other approaches used for in data minning.This paper discusses the Bayesian network and its superiority in data minning.

关 键 词：贝叶斯网络 数据挖掘技术 贝叶斯推理 贝叶斯学习 

分 类 号：TP311.131[自动化与计算机技术—计算机软件与理论] O211[自动化与计算机技术—计算机科学与技术]