A Bayesian Fuzzy Classifier for the Predisposition of Subjects to Narcotics

Abstract

A natural relationship between the probabilistic and fuzzy set approaches to the intelligent analysis of the level of risk of drug dependence was demonstrated using results of psychological testing of students. A method of forming a Bayesian model of linguistic variables characterizing the initial signs of the predisposition to drugs is presented. Steps of fuzzy logical inference of the aggregate informative classifier for ranking and classifying subjects’ psychoemotional status are considered within the framework of hierarchy analysis and principal components analysis.