id |
ddssar9611 |
authors |
de Gelder, Johan and Lucardie, Larry |
year |
1996 |
title |
Criteria for the Selection of Conceptual Modelling Languages for Knowledge Based Systems |
source |
Timmermans, Harry (Ed.), Third Design and Decision Support Systems in Architecture and Urban Planning - Part one: Architecture Proceedings (Spa, Belgium), August 18-21, 1996 |
summary |
In recent years knowledge is increasingly recognised as a critical production factor for organisations. Performance of activities such as designing, diagnosing, advising and decision making, depend on the availability and accessibility of knowledge. However, the increasing volume and complexity of knowledge endangers its availability and accessibility. By their knowledge processing competence, knowledge based systems containing a structured and explicit representation of knowledge, are expected to solve this problem. In the realisation of a knowledge based system, the phase in which a knowledge model is reconstructed through a conceptual language, is essential. Because the knowledge model has to be an adequate reflection of real-world knowledge, the conceptual language should not only offer sufficient expressiveness for unambiguous knowledge representation, but also provide facilities to validate knowledge on correctness, completeness and consistency. Furthermore, the language should supply facilities to be processed by a computer. This paper discusses fundamental criteria to select a conceptual language for modelling the knowledge of a knowledge based system. It substantiates the claim that the selection depends on the nature of the knowledge in the application domain. By analysing the nature of knowledge using the theory of functional object-types, a framework to compare, evaluate and select a conceptual language is presented. To illustrate the selection process, the paper describes the choice of a conceptual language of a knowledge based system for checking office buildings on fire-safety demands. In this application domain, the language formed by decision tables has been selected to develop the conceptual model. The paper provides an in-depth motivation why decision tables form the best language to model the knowledge in this case. |
series |
DDSS |
full text |
file.pdf (1,887,079 bytes) |
references |
Content-type: text/plain
|
last changed |
2003/08/07 16:36 |
|