计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (29): 14-16.

• 博士论坛 • 上一篇    下一篇

一种基于用户的LNS文件预测模型

刘爱贵1,2,陈 刚1   

  1. 1.中国科学院 高能物理研究所 计算中心,北京 100049
    2.中国科学院 研究生院,北京 100049
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-10-11 发布日期:2007-10-11
  • 通讯作者: 刘爱贵

User-based LAST N Successors file prediction model

LIU Ai-gui1,2,CHEN Gang1   

  1. 1.Computing Center,Institute of High Energy Physics,CAS,Beijing 100049,China 2.Graduate University of Chinese Academy of Sciences,Beijing 100049,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-10-11 Published:2007-10-11
  • Contact: LIU Ai-gui

摘要: 许多系统把数据访问请求当作是独立的事件。实际上,数据请求并非完全随机,而是由用户或程序的行为驱动的,不同的用户或程序存在不同的访问模式。LS(Last Successor)模型简单,但非常有效,然而它的预测结果严重依赖于用户或程序的访问顺序。提出了ULNS(User-based Last N Successors)文件预测模型,利用用户信息来提高预测精确度,并综合LS模型来改进算法的可适用度。实验结果表明,该预测模型具有较好的整体性能。

关键词: 文件访问模式, 预测模型, 预测精确度, 可适用度

Abstract: Most systems treat each data request as an independent event.In fact,such requests are driven by users or programs behavior,and are therefore far from random.There are different access patterns with different users or programs.LS(Last Successor)model is simple,but very available.However,its predictive results strongly depend on access order of users or programs.This paper presents ULNS(User-based Last N Successors) file prediction model,which utilizes user information to improve its accuracy.And also,it synthesizes LS model to extend its applicability.Experimental results show that the proposed prediction model has better overall performances.

Key words: file access pattern, prediction model, predictive accuracy, applicability