An application of variable-valued logic to inductive learning of plant disease diagnostic rules
R.L. Chilausky, Ben Jacobsen, Ryszard S. Michalski · International Symposium on Multiple-Valued Logic · 1976
Knowledge acquisition and representation in machines is growing in importance to computer science research. Most of the work in this direction has used traditional binary logic as a theoretical framework. However, a more adequate treatment of these problems may be achieved by an advance into multiple-valued logic. In particular, the concepts of variable-valued logic (Michalski 72,73,74a) provide useful and flexible tools for representation and automated acquisition of knowledge. This paper describes an application of AQVAL/1, a set of programs implementing the variable-valued logic system VL1, to determine diagnostic rules for soybean diseases through an inductive inference process. The methods of application and the results of an experiment are briefly described.