不确定性数据中基于GSO优化MF的模糊关联规则挖掘方法  

Fuzzy association rules mining method based on GSO optimization MF in uncertainty data

作　　者：章武媚 董琼[1] Zhang Wumei;Dong Qiong(State University of New York,Oswego,New York 13126,USA;Zhejiang Tongji Vocational College of Science & Technology,Hangzhou 310023,China)

机构地区：[1]美国纽约州立大学,美国纽约州奥斯威戈13126 [2]浙江同济科技职业学院,杭州310023

出　　处：《计算机应用研究》2019年第8期2284-2288,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(71561018);浙江省教改项目(jg20160405);全国教育信息技术研究项目(166223123)

摘　　要：针对不确定性数据中模糊关联规则的挖掘问题,提出一种基于群搜索优化(GSO)算法优化隶属度函数(MF)的模糊关联规则挖掘方法。首先,将不确定性数据通过三元语言表示模型进行表示;然后,给定一个初始MF,并以最大化模糊项集支持度和语义可解释性作为适应度函数,通过GSO算法的优化学习获得最佳MF;最后,根据获得的最佳MF,利用改进型的FFP-growth算法来从不确定数据中挖掘模糊关联规则。实验结果表明,该方法能够根据数据集自适应优化MF,以此实现从不确定数据中有效地挖掘关联规则。In order to solve the problem of mining fuzzy association rules in uncertainty data, this paper proposed a new method of mining fuzzy association rules based on optimization of membership function (MF) by group search optimization (GSO) algorithm.Firstly, it represented the uncertainty data by the 3-tuples linguistic representation model.Then, given an initial MF, it obtained the best MF by optimizing learning of GSO algorithm with maximum support of fuzzy item sets and semantic interpretability as a fitness function. Finally, it used the improved FFP-growth algorithm to mine the fuzzy association rules from the uncertain data according to the best MF obtained. Experimental results show that this method can adaptively optimize MF based on data set, so as to effectively mine association rules from uncertain data.

关 键 词：模糊关联规则挖掘 不确定数据 隶属度函数 群搜索优化算法 FFP-growth算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]