SEMANTIC MODELING OF SUBJECT DOMAINS USING PRECEDENT AND BLURRY (FUZZY) MODELS
摘要
We propose an approach to semantic modeling based on the four-level framework for knowledge representation. The method constructs a precedent model at the empirical level using precedent knowledge and then fuzzifies it to obtain statistical knowledge at the fourth level. We introduce the mathematical framework for reconstructing a set of precedents from subjective expert assessments and and outlines the characteristics of blurry models. Bibliography : 9 titles.