摘要
CILP是关系数据挖掘的主要技术之一。为提高CILP系统的效率,提出了一种基于C3模型,元学习技术和主从式静态负载平衡策略的分布式并行CILP算法,并实现了一个基于COW机群结构的分布式并行CILP原型系统。实验表明该算法是高效的,能获得较好的负载平衡,较高的加速比和并行效率。
CILP is one of the major techniques of RDM. To solve the efficiency problem in the CILP system, a distributed parallelized CILP algorithm is presented, which is based on the C^3 Model, Meta-Learn Technology and Master-Slave model of the static load balancing strategy. Also, a prototype of the distributed parallel CILP system is implemented with the COW clusters. Experiment shows that this algorithm is efficient, and can get good load balancing, speedup and parallel efficiency.
出处
《计算机应用研究》
CSCD
北大核心
2005年第9期34-36,45,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(60173014)
北京市自然科学基金资助项目(4022003)
北京工业大学研究生科技基金资助项目(YKJ-2003-45)
