首页 | 本学科首页   官方微博 | 高级检索  
     检索      

A Text Categorization System with Soft Real-Time Guarantee
作者姓名:WANG  Hua-yong  CHEN  Yu  DAI  Yi-qi
作者单位:Department of Computer Science and Technology,Tsinghua University, Beijing 100084, China
基金项目:Supported by the National Natural Science Foundation of China (90104032), the National High-Tech Research and Development Plan of China (2003AA1Z2090)
摘    要:In order to provide predictable runtime performante for text categorization (TC) systems, an innovative system design method is proposed for soft real time TC systems. An analyzable mathematical model is established to approximately describe the nonlinear and time-varying TC systems. According to this mathematical model, the feedback control theory is adopted to prove the system's stableness and zero steady state error. The experiments result shows that the error of deadline satisfied ratio in the system is kept within 4 of the desired value. And the number of classifiers can be dynamically adjusted by the system itself to save the computa tion resources. The proposed methodology enables the theo retical analysis and evaluation to the TC systems, leading to a high-quality and low cost implementation approach.

关 键 词:文本分类系统  信息检索  实时系统  反馈控制理论
文章编号:1007-1202(2006)01-0226-04
收稿时间:2005-04-28

A text categorization system with soft real-time guarantee
WANG Hua-yong CHEN Yu DAI Yi-qi.A Text Categorization System with Soft Real-Time Guarantee[J].Wuhan University Journal of Natural Sciences,2006,11(1):226-229.
Authors:Wang Hua-yong  Chen Yu  Dai Yi-qi
Institution:(1) Department of Computer Science and Technology, Tsinghua University, 100084 Beijing, China
Abstract:In order to provide predictable runtime performance for text categorization (TC) systems, an innovative system design method is proposed for soft real-time TC systems. An analyzable mathematical model is established to approximately describe the nonlinear and time-varying TC systems. According to this mathematical model, the feedback control theory is adopted to prove the system's stableness and zero steady state error. The experiments result shows that the error of deadline satisfied ratio in the system is kept within 4% of the desired value. And the number of classifiers can be dynamically adjusted by the system itself to save the computation resources. The proposed methodology enables the theoretical analysis and evaluation to the TC systems, leading to a high-quality and low-cost implementation approach.
Keywords:information retrieval  text categorization  soft real-time system  feedback control theory
本文献已被 CNKI 维普 万方数据 SpringerLink 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