Imprecise probability through f-probability and its statistical physical implications

Abstract

Abstract In this work, we present an approach to describe imprecise probability through an effective probability theory, called the f-probability. We develop a bijective and monotonous map from the precise probability in order to construct the f-probability theory based on the f-addition, f-subtraction, f-multiplication and f-division. We apply the f-probability to the Bernoulli trial and derive the f-binomial distribution. Finally, we obtain the non-extensive entropy through the f-probability theory, and give its statistical physical implications on several areas of potential applications. PACS number(s): 02.50.Cw, 05.20.-y; 05.90.+m