摘要
提出基于矩阵和权重的一种改进算法——MW_Apriori算法.该算法首先通过扫描一次事务数据库来构造0-1事务矩阵,其次赋予项和事务权重,并计算项的权重支持度,从而得到频繁项集.实验结果表明,MW_Apriori算法避免了数据库的重复扫描,使得时间和空间的耗费显著减少,同时能有效地挖掘出隐藏的、更有价值的事件.
An improved algorithm,which is called MW_Apriori algorithm is proposed based on the matrix and weight in this paper.Firstly,build the 0-1transaction matrix by scanning transaction database.Then items and transactions are assigned to weights,and the weighted support of items are caculated,accordingly gettig the frequent itemsets.Experiments show that MW_Apriori algorithm avoids rescanning the database,making the cost of time and space significantly reduced,at the same time can effectively mine the hidden and valuable rare events.
出处
《微电子学与计算机》
CSCD
北大核心
2017年第1期136-140,共5页
Microelectronics & Computer
基金
国家自然科学基金项目(61073196
61272458)
陕西省自然科学基础研究计划项目(2014JM2-6119)
榆林科技计划项目(2014CXY-12)