首页 | 本学科首页   官方微博 | 高级检索  
     

基于分而治之策略的稀疏张量多层次数据挖掘
引用本文:周溜溜,业宁,徐昇,严敏利,孙伟.基于分而治之策略的稀疏张量多层次数据挖掘[J].微电子学与计算机,2011,28(10).
作者姓名:周溜溜  业宁  徐昇  严敏利  孙伟
作者单位:南京林业大学信息技术学院,江苏南京,210037
基金项目:江苏省自然基金项目(BK2009393); 国家自然基金(30671639); 江苏省青蓝工程学术带头人项目
摘    要:分析了MET算法的局限性以及out-of-core方法的特点,融合了两种算法的思想提出一种基于分而治之策略的多层次数据挖掘算法(DRMET),避免了计算过程中可能造成的维数灾难问题,克服了MET算法执行效率不高的缺陷,同时继承了MET内存开销小的优点;实验结果表明:新算法在不增加存储空间的前提下大大约减了MET的时间开销,其效率大约是MET的1.86~15.85倍.

关 键 词:分而治之  多层次数据挖掘  DRMET  时间开销

Multi-aspect Data Minizing on the Sparse Tensor Based on Dividing and Ruling
ZHOU Liu-liu,YE Ning,XU Sheng,YAN Min-li,SUN Wei.Multi-aspect Data Minizing on the Sparse Tensor Based on Dividing and Ruling[J].Microelectronics & Computer,2011,28(10).
Authors:ZHOU Liu-liu  YE Ning  XU Sheng  YAN Min-li  SUN Wei
Affiliation:ZHOU Liu-liu,YE Ning,XU Sheng,YAN Min-li,SUN Wei(College of Computer Science and Technology,Nanjing Forestry University,Nanjing 210037,China)
Abstract:Analyzed the limitations of MET algorithm and feature of out-of-core,mixed ideas of the two algorithms and then giving a method called DRMET which based on the strategy of dividing and ruling for multi-aspects data mining;avoid the problem of dimension disaster while overcomed the MET's limitation of low efficincy,and inherit its advantage on small memory consumed;experiment gives the result: new algorithm reduced time cost from MET substantially,its efficiency improved about 1.86~15.85 times compared to ME...
Keywords:dividing and ruling  multi-aspect data mining  DRMET  time cost  
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号