Belief Functions versus Probability Functions
暂无分享,去创建一个
Two models are proposed to quantify someone's degree of belief, based respectively on probability functions, the Bayesian model, and on belief functions, the transferable belief model (Shafer 1976). The first, and by far the oldest, is well established and supported by excellent axiomatic and behaviour arguments. The model based on belief functions is often understood as some kind of generalization either of the Bayesian model or of the upper and lower probabilities model. Therefore we present our interpretation of the model developed initially by Shafer in his book (1977) and called here the transferable belief model. The major point of our interpretation is the fact-we try to dissociate completely the transferable belief model from any model based on probability functions.
[1] Cedric A. B. Smith,et al. Consistency in Statistical Inference and Decision , 1961 .
[2] Glenn Shafer,et al. A Mathematical Theory of Evidence , 2020, A Mathematical Theory of Evidence.
[3] Glenn Shafer,et al. Probability Judgment in Artificial Intelligence and Expert Systems , 1987 .
[4] Philippe Smets. Non-standard logics for automated reasoning , 1988 .