MODELING PARTIAL IGNORANCE IN ARTIFICAL INTELLIGENCE APPLICATIONS

This study aims to extend and deepen a survey of modern extensions of probability theory represented in [6, 7]. The classical probability theory possesses rather limited possibilities and cannot cope with situations of partial ignorance. Other approaches are required allowing one to solve tasks of that kind The given paper considers the generalised approach to modelling partial ignorance and its interpretation in the terms of upper and lower probabilities, second order probabilities and belief functions.