基于约束网络的因果关联规则挖掘研究

发布时间：2018-01-06 02:09

本文关键词：基于约束网络的因果关联规则挖掘研究　出处：《计算机科学》2016年S2期 　论文类型：期刊论文

【摘要】：因果关联规则是知识库中一类特殊且重要的知识类型,相对一般关联规则,其优势在于能够揭示深层知识。首先对因果关系的特征和因果关联规则的挖掘方法进行了简介。针对在挖掘初始阶段如何限定可能导致结果的原因变量集合这一问题,运用了约束网络原理来构建一个实际系统变量间的因果关系结构。通过该因果关系结构可以比较容易地导出原因变量集合及各变量的类型,从而降低挖掘的复杂性,为提高挖掘结果的准确性提供有利条件。约束网络的引入优化了因果关联规则的挖掘过程,使之趋于更完备。
[Abstract]:Causal association rules are a kind of special and important knowledge types in knowledge base, which are relative to general association rules. Its advantage lies in its ability to reveal deep knowledge. Firstly, the characteristics of causality and the mining method of causal association rules are introduced. How to define the set of cause variables that may lead to the result in the initial stage of mining is introduced. The problem. The causality structure of a real system variable is constructed by using the principle of constraint network, through which the set of cause variables and the types of variables can be easily derived. In order to reduce the complexity of mining and provide favorable conditions for improving the accuracy of mining results, the introduction of constraint networks optimizes the mining process of causal association rules and makes them more complete.
【作者单位】：中国劳动关系学院数学与计算机部;江西师范大学计算机信息工程学院;
【基金】：国家自然科学基金资助项目(61202209) 江西省教育厅科技项目(GJJ13229)资助
【分类号】：TP311.13
【正文快照】： 1引言关联规则是数据挖掘中最为成熟的领域之一。但一般的关联规则也存在不足之处,即规则更多体现的是“联系”而非“因果”。因果关联规则就是指规则的前件与后件之间存在因果关系,由于“因”的出现而导致“果”的发生。因果关联规则具有两个特点:1)规则的前件与后件之间不但

【相似文献】