Learning Probabilistic Description Logic Concepts Under Alternative Assumptions on Incompleteness.
Pasquale Minervini, Claudia d’Amato, Nicola Fanizzi, Floriana Esposito · 2014
Real-world knowledge often involves various degrees of uncertainty. For such a reason, in the Semantic Web context, difficulties arise when modeling real-world domains using only purely logical formalisms. Alternative approaches almost always assume the availability of probabilistically-enriched knowledge, while this is hardly known in advance. In addition, purely deductive exact inference may be infeasible for Web-scale ontological knowledge bases, and does not exploit statistical regularities in data. Approximate deductive and inductive inferences were proposed to alleviate such problems. This article proposes casting the concept-membership prediction problem predicting whether an individual in a Description Logic knowledge base is a member of a concept as estimating a conditional probability distribution which models the posterior probability of the aforementioned individual's concept-membership given the knowledge that can be entailed from the knowledge base regarding the individual. Specifically, we model such posterior probability distribution as a generative, discriminatively structured, Bayesian network, using the individual's concept-membership w.r.t. a set of feature concepts standing for the available knowledge about such individual.