基于正、负关联规则的数据挖掘探讨

摘要

传统的关联规则只关注于挖掘出项集间的正关联规则，但在实际应用中负关联规则同样隐含着有价值的信息。本文首先给出了正、负关联规则的定义及支持度和置信度的函数表示，重点分析了关联规则中“支持度-置信度”架构的局限性，提出了利用项集的相关性来解决关联规则中正、负矛盾规则出现的问题，同时针对置信度的设置进行了研究分析，最后对负关联规则挖掘的算法进行了讨论，旨在为关联规则的研究奠定基础。%The positive association rules only focus on mining appear between sets,but negative association rules also contains valuable information in the actual application. This article first gave the definition of the positive and negative association rules and function,focused on the analysis of association rules in the“ support confidence framework”limitation,proposed by using the correlation itemsets to solve the association rules in the positive, negative contradiction rule problems,at the same time the needle for confidence level the set was analyzed,finally the negative association rules mining algorithms were discussed,aimed at laying a foundation for the study of association rules.