基于约简数据集的关联规则挖掘策略

Association Rules Mining Strategy Based on Reduced Data Set

原始数据集中含有大量噪声数据,且数据的规模很大,直接进行关联规则挖掘会影响准确度和效率。文章提出了一种对原始数据先进行聚类,再提取关联规则的挖掘策略,可以在一定程度内减少噪声数据的干扰,消除数据对象中的冗余属性,提高规则挖掘的有效性。

There are a lot of noise data in initial data set,and the scale of data is very large,so directly mining association rules from initial data set will diminish the accuracy and efficiency.A novel association rules mining strategy based on reduced data set is presented in the paper,and the clustering analysis is the first step,so the noise data can be eliminated to a certain extent,which can improve the correctness of result-rules.The second step is rules-mining, owing to the attributes reductlon,the redundant attributes can be omitted,and the association rules can be educed in the key attributes set.