多层关联分析中数值型数据概念分层算法研究

摘要

本文主要是针对在关联分析领域，多层关联分析算法中对事务集中连续数据的层次划分算法进行相关的研究与探讨。论述了不同数值型数据分层方法的利弊，同时提出了基于信息熵的数值型数据概念分层算法，验证了该算法在提高有效频繁项集数上相对于其他分层算法具有一定的优势。%This articleis mainlyrelated tothefieldofcorrelation analysis, and it focuses on dis cussing and researching continuous datatransaction-level partitioning algorithm in multi-level associationanalysis algorithms. Discussingthe pros and consofdifferentnumeric datalayered a pproach, and proposingthenumerical datahierarchicalalgorithm based on information entropyc oncept , andverifying that thealgorithmhas certain advantagesto improvethenumberofeffective frequentitems comparing tootherhierarchicalalgorithms.