Output details
11 - Computer Science and Informatics
University of Exeter
Return to search
Output 0 of 0 in the submission
Article title
Robust Autoregression: Student-t Innovations Using Variational Bayes
Type
D - Journal article
Title of journal
IEEE Transactions on Signal Processing
Article number
-
Volume number
59
Issue number
1
First page of article
48
ISSN of journal
1941-0476
Year of publication
2011
URL
-
Number of additional authors
1
Additional information
<13> This paper appears in the principal journal for computational signal processing (impact factor 2.8). Although many signals are known to be non-Gaussian, robust inference for them is hampered by the computational complexity of Markov Chain Monte Carlo sampling. This paper introduces new tractable variational Bayesian approximations when both the innovations and observational noise are heavy-tailed. Automatic relevance determination priors are shown to accurately estimate the model order as part of the inference.
Interdisciplinary
-
Cross-referral requested
-
Research group
1 - Artificial Intelligence
Citation count
6
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-
