FOURTH INTERNATIONAL SYMPOSIUM ON
IMPRECISE PROBABILITIES AND THEIR APPLICATIONS
Carnegie Mellon University
Pittsburgh, PA, USA
July 20-23 2005

ISIPTA'05 ELECTRONIC PROCEEDINGS

Alberto Piatti, Marco Zaffalon, Fabio Trojani

Limits of Learning from Imperfect Observations under Prior Ignorance: the Case of the Imprecise Dirichlet Model

Abstract

Consider a relaxed multinomial setup, in which there may be mistakes in observing the outcomes of the process--this is often the case in real applications. What can we say about the next outcome if we start learning about the process in conditions of prior ignorance? To answer this question we extend the imprecise Dirichlet model to the case of imperfect observations and we focus on posterior predictive probabilities for the next outcome. The results are very surprising: the posterior predictive probabilities are vacuous, irrespectively of the amount of observations we do, and however small is the probability of doing mistakes. In other words, the imprecise Dirichlet model cannot help us to learn from data when the observational mechanism is imperfect. This result seems to rise a serious question about the use of the imprecise Dirichlet model for practical applications, and, more generally, about the possibility to learn from imperfect observations under prior ignorance.

Keywords. Predictive Bayesian Inference, Imprecise Dirichlet Model, Vacuous Predictive Probabilities, Imperfect Observational Mechanism

Paper Download

The paper is availabe in the following formats:

Authors addresses:

Alberto Piatti
Salita S.Michele 2
6500 Bellinzona

Marco Zaffalon
Galleria 2
CH-6928 Manno
Switzerland

Fabio Trojani
Rosenbergstr. 52
CH-9000 St.Gallen

E-mail addresses:

Alberto Piatti alberto.piatti@lu.unisi.ch
Marco Zaffalon zaffalon@idsia.ch
Fabio Trojani fabio.trojani@unisg.ch


[ back to the Proceedings of ISIPTA'05 home page 
Send any remarks to the following address: smc@decsai.ugr.es