One-sided approximation of Bayes rule and its application to regression model with Cauchy noise
Miroslav Kárný, Katalin Mária Hangos · Czech digital mathematics library · 1988
The paper presents results which aim to broaden applicability of Bayesian identification to non-standard problems by using a systematic approach to the design of a feasible approximation of the optimal, but unfeasible solution.The essence of the theory, admitting to generate global approximants in real time, is outlined.The approximation task formulated as a properly chosen problem of mathematical programming is applied to the practically important case of the regression with heavy tailed noise.