摘要
在知识发现过程中 ,通过挖掘算法产生大量的模式 ,但是大多数用户对此不感兴趣。如何对它们进行评价 ,选取出用户感兴趣的和有用的知识成为至关重要的一环 ,故对知识评价的研究具有重要的意义。首先分析了评价过程与知识发现的结合方式 ;针对评价的综合度量标准 (感兴趣度 ) ,从客观性和主观性两个方面分别进行了介绍 ;并针对因果关联规则概述了我们提出的一种新的评价方法。
In KDD(Knowledge Discovery in Database),there are a number of patterns discovered from large database by data mining,but most of them are of no interestingness to the user. How to evaluate them so that the user can get those which are useful and interesting is crucial. Therefore it is of great value to discuss the evaluation for the discovered patterns. At first, the present situation of research for evaluation is given in the paper; then the standard of evaluation-measure of interestingness-is respectively analyzed from the objective and subjective view. For causal rules, a new evaluating method is generally presented
出处
《计算机应用研究》
CSCD
北大核心
2001年第12期1-4,20,共5页
Application Research of Computers
基金
国家自然科学基金重点资助项目 ( 6 9835 0 0 1 )
关键词
知识发现
数据挖掘
感兴趣度
关联规则
知识评价
数据库
Knowledge Discovery
Data Mining
Measures of Interestingness
Evaluation
Causal Rule
