SENSITIVITY OF A PRIORY LOGICAL PROBABILITY CONCLUSION IN INTELLECTUAL INFORMATIONAL SYSTEMS
SPIIRAS, Saint Petersburg, 199178, Russian Federation; Saint Petersburg State University, Saint Petersburg, 199178, Russian Federation, ; Head of Laboratory; Professor
Abstract. A method is proposed for estimation of the effect of initial data variation on an a priory conclusion in a fragment of algebraic Bayes net. The method is based on solution of a set of linear programming problems. An upper estimate is derived for conclusion sensitivity depending linearly on initial data variation radius.
Keywords: estimation of sensitivity, logical probability conclusion, algebraic Bayes net.