一种新的频繁模式挖掘算法  

New frequent patterns mining algorithm

作　　者：叶海琴[1] 廖利[1] 王意锋 张爱玲 

机构地区：[1]周口师范学院计算机科学与技术学院,河南周口466001 [2]73658部队,安徽滁州239421 [3]71352部队自动化站,河南安阳455000

出　　处：《南京理工大学学报》2016年第1期29-34,共6页Journal of Nanjing University of Science and Technology

基　　金：国家自然科学基金(U1504613);河南省软科学研究项目(142400411220);河南省科技厅基础前沿项目(142300410432);河南省高等学校重点科研项目(15B520031)

摘　　要：为了适应由于进行添加、删除、修改操作而频繁变化的数据库以及加速支持度求解过程,该文提出了一种新的频繁模式挖掘算法。该算法将顾客的一次购买行为转化为比特串,通过对比特串的操作,逐渐更新事务集的典型集,从而适应目前数据库的频繁变化。典型集中包含了所有模式,根据支持度阈值可以从典型集中快速找到频繁模式。通过实例分析了该算法面对频繁变化数据库的过程,表明了该算法具有很强的适应数据库变化的能力,并能够根据给定的支持度阈值快速求出所需的频繁模式,仿真实验验证了该算法的有效性和可行性。In order to adapt to the frequent changes of the database by adding, deleting or modifying operations and speeding up the solving process of support, this paper proposes a new frequent patterns mining algorithm. To adapt to the frequent changes in the current database, customers＇ once purchase behavior is converted into a bit string and the typical set of transaction sets is updated gradually by the operation on bit strings in this paper. The typical set includes all patterns. Frequent patterns can be found quickly from the typical set according to the support threshold. An example is used to analyse the process of the algorithm in the face of the frequent change database. It shows that the algorithm has strong ability of adapting to changes in the database and can find the frequent patterns quickly according to the given support threshold. Simulation results verify the effectiveness and feasibility of the algorithm.

关 键 词：频繁模式 挖掘算法 比特串 支持度 典型集 事务集 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]