Stime fiduciali di probabilità nell'aggregazione bayesiana di informazioni

Abstract

Among the aggregative algorithms suitable to compose various probability densities in a single synthesis distribution - which abstracts information, evaluations, opinions coming from different sources about an unknown quantity -, the bayesian scheme suggested by Peter A. Morris is characterized by the recourse to a "calibration function": a kind of probabilistic balancer, technically configurable like second order prob-ability density and consequently specifiable by fiducial models for estimating probabilities.