Ontology Based Statistical Automated Inference - New Approach to Artificial Intelligence

Abstract

Statistical analysis requires understanding the nature of the phenomenon under study, as well as understanding sense of mathematical statistics. Bridging the gap between semantic web based on knowledge representation languages, and concepts described by mathematical formula is a challenge for AI. In order to overcome this gap the ontology language P-ONT (based on directed graph) has been invented. To illustrate the capabilities of the P-ONT language, semantic web (built on the P-ONT ontology) OLAP cube, relational data bases and generalized hierarchical statistical regression models are presented.