A Statistical Pattern Recognition Method for Providing User Demand in Community Computing

커뮤니티 컴퓨팅에서 사용자 요구 반영을 위한 통계적 패턴 인식 기법

  • 김성빈 (숭실대학 컴퓨터학과) ;
  • 정혜동 (전자부품연구원 지능형 정보연구센터) ;
  • 이형수 (전자부품연구원 지능형 정보연구센터) ;
  • 김석윤 (숭실대학 컴퓨터학과)
  • Published : 2009.05.07

Abstract

The conventional computing is a centralizing system, but it has been gradually going to develop ubiquitous computing which moves roles away from the main. The Community Computing, a new paradigm, is proposed to implement environment of ubiquitous computing. In this environment, it is important to accept the user demand. Hence in this paper recognizes pattern of user's activity statistically and proposes a method of pattern estimation in community computing. In addition, user's activity varies with time and the activity has the priority We reflect these. Also, we improve accuracy of the method through Knowledge Base organization and the feedback system. We make program using Microsoft Visual C++ for evaluating performance of proposed method, then simulate it. We can confirm it from the experiment result that using proposal method is better in environment of community computing.

Keywords