基于中间件技术的数据挖掘企业应用平台的研究与实现  

Data Mining Platform for Enterprise Application Based on Middleware

作　　者：师文轩[1] 黄亚楼[1] 谢茂强[1] 

机构地区：[1]南开大学软件学院,天津300071

出　　处：《南开大学学报（自然科学版）》2009年第6期15-20,共6页Acta Scientiarum Naturalium Universitatis Nankaiensis

基　　金：天津市科技攻关计划重点科技攻关专项项目(05YFGZGX24000)

摘　　要：数据挖掘具有计算密集型和存储密集型的特点,中间件技术能够较好的解决这两个问题.研究并实现了典型的分类、聚类、关联规则算法及其增量算法的中间件和数据挖掘企业应用平台,能够处理100 Mbit量级的数据,适应的数据增量在10～100 Mbit量级,并且能够根据不同的挖掘任务实现相应的模式展现与可视化.平台上对某网球训练基地运动员体能训练数据集执行增量聚类挖掘任务,结果表明该平台能较好地满足可靠性、扩展性、易用性等业务需要.Data mining application is dense in computing and dataset storage and middle-ware can easily resolve these problems.The system realizes some classic arithmetic middleware about classification, clustering, association rules and relevant incremental arithmetic middle- ware, moreover the system realizes a data mining platform, which can deal with large dataset of 100 Mbit and incremental dataset that is between 10 and 100 Mbit. The platform can imple- ment the visualization of the results of some data mining tasks. The platform has executed an incremental clustering task using the athlete physical training dataset of a tennis training base, and the data mining result has showed a good reliability, expansibility and facility to be avail-able for the business demand.

关 键 词：中间件 增量算法 挖掘任务 可视化 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]