Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

REDER: REtrieval of Documents based on Evidential Reasoning

Lung Albert Chen

EECS Department
University of California, Berkeley
Technical Report No. UCB/CSD-90-572
June 1990

http://www.eecs.berkeley.edu/Pubs/TechRpts/1990/CSD-90-572.pdf

Knowledge-based retrieval of information has been proved to enhance the performance of conventional retrieval systems. Systems like RUBRIC (RUle-Based Retrieval of Information by Computer) and KADR (Knowledge Assisted Document Retrieval) have successfully incorporated the techniques developed in the research of knowledge-based systems to solve the retrieval problem. Current knowledge-based retrieval systems, however, share several common problems. They are designed either for special users or their reasoning mechanism is not general enough. When developed for specific users, they reflect user preference instead of expert knowledge; when the reasoning mechanism is not sufficiently general, an inadequate model could lead to counterexamples. The focus of this research is to design a theoretically sound knowledge-based retrieval system for general users. In this system, expert knowledge serves as a basis for automatic query formulation and query evaluation; furthermore, the entire retrieval of information is considered to be a process of evidential reasoning. The foundation of the reasoning methodology is Dempster-Shafer (D-S) theory and its extension. The new approach subsumes the conventional Boolean system and has many advantages over the other retrieval systems. REDER (REtrieval of Documents based on Evidential Reasoning) is a prototype of such a system and demonstrates a complete process of knowledge-based retrieval of information. The results of a retrieval experiment verifies the performance of this system.


BibTeX citation:

@techreport{Chen:CSD-90-572,
    Author = {Chen, Lung Albert},
    Title = {REDER: REtrieval of Documents based on Evidential Reasoning},
    Institution = {EECS Department, University of California, Berkeley},
    Year = {1990},
    Month = {Jun},
    URL = {http://www.eecs.berkeley.edu/Pubs/TechRpts/1990/6178.html},
    Number = {UCB/CSD-90-572},
    Abstract = {Knowledge-based retrieval of information has been proved to enhance the performance of conventional retrieval systems. Systems like RUBRIC (RUle-Based Retrieval of Information by Computer) and KADR (Knowledge Assisted Document Retrieval) have successfully incorporated the techniques developed in the research of knowledge-based systems to solve the retrieval problem. Current knowledge-based retrieval systems, however, share several common problems. They are designed either for special users or their reasoning mechanism is not general enough. When developed for specific users, they reflect user preference instead of expert knowledge; when the reasoning mechanism is not sufficiently general, an inadequate model could lead to counterexamples. The focus of this research is to design a theoretically sound knowledge-based retrieval system for general users. In this system, expert knowledge serves as a basis for automatic query formulation and query evaluation; furthermore, the entire retrieval of information is considered to be a process of evidential reasoning. The foundation of the reasoning methodology is Dempster-Shafer (D-S) theory and its extension. The new approach subsumes the conventional Boolean system and has many advantages over the other retrieval systems. REDER (REtrieval of Documents based on Evidential Reasoning) is a prototype of such a system and demonstrates a complete process of knowledge-based retrieval of information. The results of a retrieval experiment verifies the performance of this system.}
}

EndNote citation:

%0 Report
%A Chen, Lung Albert
%T REDER: REtrieval of Documents based on Evidential Reasoning
%I EECS Department, University of California, Berkeley
%D 1990
%@ UCB/CSD-90-572
%U http://www.eecs.berkeley.edu/Pubs/TechRpts/1990/6178.html
%F Chen:CSD-90-572