基于分解数据库的FP-growth算法关联规则研究  

Research on Association Rules of FP-growth Algorithm Based on Decomposition Database

作　　者：刘艺 张海涛 刘奇燕 石硕[1] LIU Yi;ZHANG Haitao;LIU Qiyan;SHI Shuo(College of Information Science and Engineering,China Ocean University,Qingdao 266100;Technical Research Center,China Tobacco Yunnan Industrial Co.,Ltd,Kunming 650024)

机构地区：[1]中国海洋大学信息科学与工程学院,青岛266100 [2]云南中烟工业有限责任公司技术中心,昆明650024

出　　处：《计算机与数字工程》2018年第7期1306-1310,1416,共6页Computer & Digital Engineering

基　　金：国家重点研发计划(编号:2016YFB1001103)资助

摘　　要：论文针对频繁模式增长算法(FP-growth)中存在的频繁模式树(FP-tree)占据空间过大等问题,提出了一种改进的FP-growth算法,该算法采用分解数据库思想对事务数据库进行分类后分别挖掘以提高算法效率,并在提取规则时增加约束条件以更好地适用于所研究的医疗数据。实验结果表明,该算法的计算效率、产生的关联规则数量方面的性能明显优于经典的Apriori算法和FP-growth算法。通过对糖尿病以及它的三种主要并发症的关联规则的研究,获得糖尿病主要并发症发病概率定量关系(高血压>高脂血症>冠心病)以及肥胖增大患糖尿病并发症概率的规则,对于糖尿病并发症的前期预防有一定参考价值。In this paper,an improved FP-growth algorithm is proposed to solve the problem that the FP-tree needs large storage exists in the FP-growth algorithm.The algorithm uses the idea of decomposing the database to classify the transaction database,improve the efficiency of the algorithm.And increase the constraints in the extraction rules to better apply to the medical data studied.The experimental results show that the computational efficiency and the number of associated rules are superior to the classical Apriori algorithm and FP-growth algorithm.Through the study of the association rules of diabetes and its three major complications,we acquire the quantitative relationship between the incidence of major complications of diabetes mellitus(hypertension>hyperlipidemia>coronary heart disease)and the rule that obesity increases the probability of complications of diabetes mellitus.These conclusions have a certain reference value for the early prevention of diabetes complications.

关 键 词：改进FP-growth算法 关联规则 散列表 数据库分解 规则提取 糖尿病并发症 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]