AUEB STATS SEMINARS 10/1/2019: Group invariance for objective prior elicitation in testing problems by Anabel Forte
Mon 7 Jan 2019 - 13:23
AUEB STATISTICS SEMINAR SERIES JANUARY 2019
Anabel Forte Deltell
University of Valencia
Group invariance for objective prior elicitation in testing problems
THURSDAY 10/1/2019
13:00
ΑΙΘΟΥΣΑ Τ105
NEW AUEB BUILDING, (TROIAS 2)
ABSTRACT
In a Bayesian framework, Bayes factors and posterior probabilities of hypotheses are formal tools to approach testing problems with prior elicitation being one of the main challenges. This is a particularly delicate issue in an objective scenario where the prior information is scarce.
Bayarri et al. (2012) introduced a number of criteria upon which the prior elicitation of any testing problem should rely. Among those, the invariance criterion emerge as one of the most relevant. This criterion states that if the entertained models have a common group invariant structure this should be preserved after marginalization with respect to the prior. In this context, Bayarri et al. considered a general invariance group, common to any linear model but, if you look into the specific structure of a given model others transformations can be established.
This work deepens in the meaning and effect of this invariance criteria for a generalized scenario considering a model-specific type of invariance. The criterion hence, leads us to a better characterization of the prior distribution for the specific testing problem.
Facebook event: https://www.facebook.com/events/225837571638635/
Anabel Forte Deltell
University of Valencia
Group invariance for objective prior elicitation in testing problems
THURSDAY 10/1/2019
13:00
ΑΙΘΟΥΣΑ Τ105
NEW AUEB BUILDING, (TROIAS 2)
ABSTRACT
In a Bayesian framework, Bayes factors and posterior probabilities of hypotheses are formal tools to approach testing problems with prior elicitation being one of the main challenges. This is a particularly delicate issue in an objective scenario where the prior information is scarce.
Bayarri et al. (2012) introduced a number of criteria upon which the prior elicitation of any testing problem should rely. Among those, the invariance criterion emerge as one of the most relevant. This criterion states that if the entertained models have a common group invariant structure this should be preserved after marginalization with respect to the prior. In this context, Bayarri et al. considered a general invariance group, common to any linear model but, if you look into the specific structure of a given model others transformations can be established.
This work deepens in the meaning and effect of this invariance criteria for a generalized scenario considering a model-specific type of invariance. The criterion hence, leads us to a better characterization of the prior distribution for the specific testing problem.
Facebook event: https://www.facebook.com/events/225837571638635/
- AUEB STATS SEMINARS 20/12/2019: Latent group structure and regularized regression by Konstantinos Perrakis
- STATS AUEB Seminars 20/12/2017: Analytics within the European Business World: An AUEB Stats Graduate sharing his experiences by Panagiotis Pramangioulis (ECKART GmbH, ALTANA Group, Germany)
- AUEB Stats PhD Seminars 2019 Part B: Updated Schedule for September-December 2019
- AUEB Stats PhD Seminars 2019: Week 4
- AUEB Stats PhD Seminars 2019: Week 8
Permissions in this forum:
You cannot reply to topics in this forum