A constraint-based approach to uncertain and imprecise reasoning. Application to expert systems
The methodology by constraints seems to be very general. It is a way to define deduction methods based on the search of particular results (bottom up method) instead of purely theoretical aspects whose results are subsequently verified (top down).
An application to expert rules has been presented. It is worthnoting that we could do the most exhaustive exploration as possible (by using all the known formulations of rules and all the “interesting” constraints) and, so, make an inventory of the advantages and disadvantages (which depends on their use) of each type of rules for each constraint, and provide, in this way, a guide for the choice of methods.
Unable to display preview. Download preview PDF.
- Dubois, D. and Prade, H. A typology of fuzzy “if... then...” rules. In: Proceedings of the congress the coming of age of fuzzy logic (3rd IFSA congress). Washington: 1989. p. 782–785Google Scholar
- Dubois, D. and Prade, H. Gradual inference rules in approximate reasoning. In: Advances in approximate reasoning based on fuzzy logic. Toulouse: IRIT, 1990. 54 p. Rapport IRIT/90-6/RGoogle Scholar