Acquisition and Representation of Knowledge for the FOREX Expert System

Abstract

In the FIW II project, an expert system shall be developed for the management of forestries and especially for the rehabilitation of forest ecosystems dominated by Norway spruce. From the artificial intelligence point of view, the domain is characterised by its large number of partly contradictory and vague theories supplied by locally distributed experts. In this paper, the required knowledge is analysed and knowledge representation formalisms are proposed for this domain. It is tried to cover the whole domain and not only the rehabilitation problem to enhance extensions of the knowledge base to be conducted in the future. An expert system as well as an Internet browser has access to the knowledge base on a web server. The knowledge acquisition is supported by the Internet browser. A first prototype of a knowledge base for the bark-beetle problem was developed with PROLOG (Krzizala 1997). A first knowledge acquisition tool on base of HTML were developed in a second master thesis (Ledl 1996). This paper describes these systems, summarise first results and makes some proposal for further improvement.