Volume 13 Part 2 Article 51
Title: Development of an Expert System for the Cultivation of Mushrooms and Validation of the Prototype
Authors: J.W.M. Backus, A.J.I. Roestel and L.J.L.D. van Griensven
The aim of this research project is to develop a way of making knowledge available for mushroom growers by using expert system technology. Two problems are subject of discussion: How to represent knowledge independent from a programming language? Which factors are influencing system performance? The problem solving strategy of the expert is analyzed to be a heuristic classification method. Concepts in different classification categories are related by non-hierarchical inferences. Validation of a prototype showed the performance of the system. In certain areas, knowledge refinement is necessary. Differences between experts in solving problems in the domain are shown. The use of well-defined, objective measurable data increases the ability of the system to produce solutions. As specific domain of knowledge, we have chosen bacterial blotch disease, caused by Pseudomonas tolaasii. The dialogue between expert system and user is supported by images, stored on a laser vision disc.Please login to download the PDF for this proceeding.