A Method of Probability Transformation Based on the Assignable Certainty

To deal with the uncertainty problems by using Dempster-Shafer (D-S) evidence theory, the reliability values of the various propositions usually need to be transformed into the basic probabilities, which facilitates a correct decision-making. However, some probability transformation algorithms may not make full use of the prior information and bring about risks of the wrong decision due to the inaccurate probability transformation. Thus, this paper proposes an algorithm of the probability transformation based on the Assignable Certainty, in which different transformation rules are respectively presented to handle two cases of all existence/partial existence of single proposition. Numerical experiments demonstrate that the proposed algorithm is more objective and accelerates the convergence process. It not only improves the accuracy of decision-making, but also reduces the risk of decision-making.