| 131 | 3 | 46 |
| 下载次数 | 被引频次 | 阅读次数 |
通过对Apriori和1-k-Apriori两种算法进行细致分析和深入研究,结合这两种算法的优点,提出了I1-k-Apriori算法.1-k-Apriori算法中利用Lk-1与L1相连接来得到候选项目集,但是,有些情况下,用这种方法生成的候选k项集数量过大,导致k项集的筛选代价太大.I1-k-Apriori算法根据k-1项集的特性和事务数据库中数据的特性来决定产生k项集的方法,可以有效避免由于Lk项数过多而影响运算效率.实验结果表明,I1-k-Apriori算法较大提高了运算效率.
Abstract：The Apriori algorithm and the 1-k-Apriori algorithm are analyzed.I1-k-Apriori algorithm is proposed,in which the advantages of the above mentioned algorithms are used.The idea of 1-k-Aprior is to generate the Lk by combining Lk-1 and L1.However,the number of candidate Lk item in 1-k-Apriori may be far more than that in Apriori in some cases.I1-k-Apriori chooses the method of Lk generating according to the feature of the Lk-1 and the size of transaction database so that the over cost can be avoided.The results of experiments show that I1-k-Apriori algorithm is more efficient.
[1]Han J W,Kamber M.数据挖掘概念与技术[M].范明,译.北京:机械工业出版社,2005.
[2]Perego R,Orlando S,Palmerini P.Enhancing the Apriori algorithm for frequent set counting[C]∥Proceedings of the 3rdInternational Conference on Data Warehousing and Knowledge Discovery.London:Springer-Verlag,2001:71-82.
[3]Jovanoski V,Lavrac N.Classification rule learning with Apriori-C[C]∥Proceedings of the 10th Portuguese Conferenceon Artificial Intelligence on Progress in Artificial Intelligence,Knowledge Extraction,Multi-agent Systems,Logic Pro-gramming and Constraint Solving.London:Springer-Verlag,2001:44-51.
[4]钱雪忠,孔芳.关联规则挖掘中对Apriori算法的研究[J].计算机工程与应用,2008,44(17):138-140.
[5]郭健美,宋顺林,李世松.基于Apriori算法的改进算法[J].计算机工程与设计,2008,29(11):2814-2815.
[6]何中胜,庄燕滨.基于Apriori&Fp-growth的频繁项集发现算法[J].计算机技术与发展,2008,18(7):45-47.
[7]张倩,王致和,杨俊.一种Apriori的改进算法[J].沈阳理工大学学报,2006,25(1):40-42.
基本信息:
中图分类号:TP311.13
引用信息:
[1]綦孝姬,于红,刘溪婧,等.基于候选项目集特性的改进Apriori算法研究[J].郑州大学学报(理学版),2009,41(01):36-39.
基金信息:
大连市青年基金资助项目,编号2005J22JH038;; 辽宁省教育厅基金资助项目,编号05L090
2009-03-15
2009-03-15