The Method for Automated Determination of Certainty Coefficients Values of Initial Facts in Decision-making Systems Based on Fuzzy Approximate Reasoning Networks

The article considers the approach for automated determining the values of confidence coefficients to the initial facts in approximate reasoning networks. The proposed method is based on construction of membership functions for each coefficient and using of the mathematical apparatus of fuzzy logic to fine-tune the initial rough estimate of the coefficient value given at design stage of an intelligent decision-making system. The developed method makes it possible to automate the procedure for determining the exact values of coefficients as much as possible, reducing the need for their a priori expert assessment and allowing the approximate reasoning network to correspond as closely as possible to the conditions of a particular real situation in which a decision is made.