Hierarchical Bayesian ARX models for robust inference

Title: Hierarchical Bayesian ARX models for robust inference
Creator: Dahlin, Johan; Lindsten, Fredrik; Schon, Thomas B.; Wills, Adrian
Relation: 16th IFAC Symposium on System Identification 2012. Proceedings of IFAC 2012: 16th IFAC Symposium on System Identification (Bruxelles, Belgium 11-13 July, 2012) p. 131-136
Publisher Link: http://dx.doi.org/10.3182/20120711-3-BE-2027.00318
Publisher: IFAC
Resource Type: conference paper
Date: 2012
Description: Gaussian innovations are the typical choice in most ARX models but using other distributions such as the Student's t could be useful. We demonstrate that this choice of distribution for the innovations provides an increased robustness to data anomalies, such as outliers and missing observations. We consider these models in a Bayesian setting and perform inference using numerical procedures based on Markov Chain Monte Carlo methods. These models include automatic order determination by two alternative methods, based on a parametric model order and a sparseness prior, respectively. The methods and the advantage of our choice of innovations are illustrated in three numerical studies using both simulated data and real EEG data.
Subject: ARX models; robust estimation; Bayesian models; Markov Chain Monte Carlo
Identifier: http://hdl.handle.net/1959.13/1309112
Identifier: uon:21778
Identifier: ISBN:9783902823069
Language: eng
Reviewed

		Thumbnail	File	Description	Size	Format