基于大数据集的动态数据库关联挖掘研究  

Research on Dynamic Database Association Mining Based on Big Data Set

作　　者：刘张榕[1] LIU Zhangrong(Information Engineering Department, Fujian Forestry Vocational Technical College, Nanping 353000, China)

机构地区：[1]福建林业职业技术学院信息工程系,福建南平353000

出　　处：《微型电脑应用》2021年第12期60-62,67,共4页Microcomputer Applications

基　　金：2020年福建省中青年教师教育科研项目(科技类)立项项目(JAT201119)。

摘　　要：动态数据库关联挖掘方法存在挖掘时运行时间长、内存使用量大、各节点任务分布不均等问题,因而提出了基于大数据集的动态数据库关联挖掘研究。采用大数据集中分布式计算,分布式存储动态数据库中的历史数据;针对其存储的数据序列,进行数据修剪重排分组和计算量预估与均衡化分组,形成动态数据库历史数据;更新动态数据库新增数据,分别挖掘动态数据库中的历史数据和新增数据,并将数据合并,完成动态数据库关联挖掘。结果表明,采用该方法进行动态数据库关联挖掘,受到支持度影响较小,运行速度快,对内存使用量低,能均匀分配各节点的计算量。Aiming at the problem of long running time,large memory usage,and uneven distribution of tasks among nodes in the dynamic database association mining method,a research on dynamic database association mining based on large data sets is carried out.Using big data centralized distributed computing,historical data are stored distributively in the dynamic database;according to the stored data sequence,data are pruned and rearranged to form new groups which are estimation and equalization groups,the historical data form dynamic database.The dynamic database is updated by adding new data,historical data and new data are mined,respectively,and are merged to complete the dynamic database association mining.The results show that the use of this method for dynamic database association mining is less affected by the degree of support,fast running,low memory usage,and can evenly distribute the calculation amount of each node.

关 键 词：大数据集 动态数据库 数据库关联 关联挖掘 

分 类 号：TP301[自动化与计算机技术—计算机系统结构] TP311.13[自动化与计算机技术—计算机科学与技术]