Rule Refinement Using the Probabilistic Rule Generator

This work treats the case of expert-originated hypotheses which are to be modified or refined by training event data. The method accepts the hypotheses in the form of weighted VL expressions and uses the probabilistic rule generator, PRG. The theory of operation, verified by experimental results, provides for any degree of hypothesis modification, ranging from minor perturbation to complete replacement according to supplied confidence weightings.

Authors

• Won D. Lee Sylvian Ray

Keywords

No keywords are indexed for this paper.