基于相关性度量的关联规则挖掘

摘要

The traditional method of data mining based on association rules can easily bring about some useless and e-ven misleading rules, and it also fails to differentiate between the positive association rules and the negative ones. This paper aims to solve these problems. First the interest model is built up on the basis of the correlation measure, then by utilizing this model the algorithm of association rules is greatly improved, and lastly experimental results have proved that this algorithm can avoid useless and misleading rules and generate some interesting negative association rules.%传统的关联规则挖掘方法客易生成一些无用规则,甚至生成误导规则,也不能区分正负关联规则.通过相关性度量,建立了基于相关性度量的兴趣度模型,并利用兴趣度模型改进了关联规则算法,最后,通过实例验证了此算法不仅能够避免生成无用规则和误导规则,还能生成一些感兴趣的负关联规则.