Filters








82 Hits in 6.4 sec

Combining a relaxed EM algorithm with Occam's razor for Bayesian variable selection in high-dimensional regression

Pierre Latouche, Pierre-Alexandre Mattei, Charles Bouveyron, Julien Chiquet
2016 Journal of Multivariate Analysis  
We address the problem of Bayesian variable selection for high-dimensional linear regression.  ...  Model selection is performed afterwards relying on Occam's razor and on a path of models found by the EM algorithm.  ...  In Section 5, a new algorithm, called "spinyReg", for variable selection in high-dimensional regression is introduced.  ... 
doi:10.1016/j.jmva.2015.09.004 fatcat:pg2gwwcwhna5pfmu5wnzyvg4s4

Automatic image orientation detection

A. Vailaya, Hong Jiang Zhang, A. Jain
1999 Proceedings 1999 International Conference on Image Processing (Cat. 99CH36348)  
We present an algorithm for automatic image orientation estimation using a Bayesian learning framework.  ...  We further show how principal component analysis (PCA) and linear discriminant analysis (LDA) can be used as a feature extraction mechanism to remove redundancies in the high-dimensional feature vectors  ...  Due to this ability of nonparametric approaches to scale well to high-dimensional data, we have selected LVQ as the VQ algorithm in our approach. A.  ... 
doi:10.1109/icip.1999.822965 dblp:conf/icip/VailayaZJ99 fatcat:ofcot5lktrcijdvkgchjemn72q

Automatic image orientation detection

A. Vailaya, H. Zhang, HongJiang Zhang, Feng-I Liu, A.K. Jain
2002 IEEE Transactions on Image Processing  
We present an algorithm for automatic image orientation estimation using a Bayesian learning framework.  ...  We further show how principal component analysis (PCA) and linear discriminant analysis (LDA) can be used as a feature extraction mechanism to remove redundancies in the high-dimensional feature vectors  ...  Due to this ability of nonparametric approaches to scale well to high-dimensional data, we have selected LVQ as the VQ algorithm in our approach. A.  ... 
doi:10.1109/tip.2002.801590 pmid:18244671 fatcat:xhnbstyagvcpznybmassg27giq

Bayesian Variable Selection for Globally Sparse Probabilistic PCA [article]

Charles Bouveyron, Pierre Latouche, Pierre-Alexandre Mattei
2016 arXiv   pre-print
Sparse versions of principal component analysis (PCA) have imposed themselves as simple, yet powerful ways of selecting relevant features of high-dimensional data in an unsupervised manner.  ...  To avoid the drawbacks of discrete model selection, a simple relaxation of this framework is presented. It allows to find a path of models using a variational expectation-maximization algorithm.  ...  The variable x therefore converges in probability to Wy, which follows a Bessel(1/α, (d − q)/2) distribution according to our lemma.  ... 
arXiv:1605.05918v2 fatcat:sckpgvdblnbpviisplroaxgmma

Bayesian variable selection for globally sparse probabilistic PCA

Charles Bouveyron, Pierre Latouche, Pierre-Alexandre Mattei
2018 Electronic Journal of Statistics  
The variable x therefore converges in probability to Wy, which follows a Bessel(1/α, (d − q)/2) distribution according to our lemma.  ...  Since zero is a constant, this convergence also happens to be in probability (Van der Vaart, 2000, p. 10).  ...  Combined with a greedy technique similar to Occam's window (Madigan and Raftery, 1994) , this feature could allow for example to perform Bayesian model averaging, which is not possible with ARD.  ... 
doi:10.1214/18-ejs1450 fatcat:imv2bdsolbfpba3vh2h3yj4oty

A review on probabilistic graphical models in evolutionary computation

Pedro Larrañaga, Hossein Karshenas, Concha Bielza, Roberto Santana
2012 Journal of Heuristics  
Specifically, we give a survey of probabilistic model building-based evolutionary algorithms, called estimation of distribution algorithms, and compare different methods for probabilistic modeling in these  ...  Evolutionary algorithms is one such discipline that has employed probabilistic graphical models to improve the search for optimal solutions in complex problems.  ...  With a decomposable metric, the score of a Bayesian network can be computed as the combination of scores obtained for smaller factors (e.g., a single variable).  ... 
doi:10.1007/s10732-012-9208-4 fatcat:54ipbzsryfbt5nqmaczgurb2he

A survey of fuzzy clustering algorithms for pattern recognition. I

A. Baraldi, P. Blonda
1999 IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics)  
Moreover, a set of functional attributes is selected for use as dictionary entries in the comparison of clustering algorithms, which is the subject of Part II of this paper [1].  ...  From this discussion an equivalence between the concepts of fuzzy clustering and soft competitive learning in clustering algorithms is proposed as a unifying framework in the comparison of clustering systems  ...  These alternative approaches have to deal effectively with the curse of dimensionality and with the qualitative principle known as Occam's razor.  ... 
doi:10.1109/3477.809032 pmid:18252357 fatcat:caopmkf2ejeqnmbok4rmdgmn2q

A novel Hybrid RBF Neural Networks model as a forecaster

Oguz Akbilgic, Hamparsum Bozdogan, M. Erdal Balaban
2013 Statistics and computing  
HRBF-NN is a flexible forecasting technique that integrates regression trees, ridge regression, with radial basis function (RBF) neural networks (NN).  ...  We develop a new computational procedure using model selection based on information-theoretic principles as the fitness function using the genetic algorithm (GA) to carry out subset selection of best predictors  ...  We further acknowledge the valuable comments of the three anonymous referees and the Associate Editor which resulted to a much improved paper.  ... 
doi:10.1007/s11222-013-9375-7 fatcat:2rbjshe5ozgirapchxzw5xhboi

Online Bayesian tree-structured transformation of HMMs with optimal model selection for speaker adaptation

Shaojun Wang, Yunxin Zhao
2001 IEEE Transactions on Speech and Audio Processing  
To balance between model complexity and goodness of fit to adaptation data, a dynamic programming algorithm is developed for selecting models using a Bayesian variant of the "minimum description length  ...  This paper presents a new recursive Bayesian learning approach for transformation parameter estimation in speaker adaptation.  ...  Mokbel for providing their preprints. Finally, they thank J. J. Liu of Beckman Institute, University of Illinois at Urbana-Champaign, for a helpful discussion about Bayesian model selection.  ... 
doi:10.1109/89.943344 fatcat:wqakgiaxrjdcnovxeql5ulnzwq

A Bayesian approach to the mixed-effects analysis of accuracy data in repeated-measures designs

Yin Song, Farouk S. Nathoo, Michael E.J. Masson
2017 Journal of Memory and Language  
To fit this model we derive an efficient procedure for simultaneous point estimation and model selection based on the iterated conditional modes algorithm combined with local polynomial smoothing.  ...  We propose a Bayesian approach for the mixed-effects analysis of accuracy studies using mixed binomial regression models and we investigate techniques for model selection. v  ...  Farouk Nathoo for top-notch research guidance, illuminating chats, buying coffee, providing funding, sharing cookies and frequent advice over the years at University of Victoria.  ... 
doi:10.1016/j.jml.2017.05.002 fatcat:vxkalbnuwfgdhgutfpdzuvqsty

CrossCat: A Fully Bayesian Nonparametric Method for Analyzing Heterogeneous, High Dimensional Data [article]

Vikash Mansinghka, Patrick Shafto, Eric Jonas, Cap Petschulat, Max Gasner, Joshua B. Tenenbaum
2015 arXiv   pre-print
There is a widespread need for statistical methods that can analyze high-dimensional datasets with- out imposing restrictive or opaque modeling assumptions.  ...  CrossCat is based on approximately Bayesian inference in a hierarchical, nonparamet- ric model for data tables.  ...  The authors found it surprising that a reliable and scalable implementation was possible. Several authors were involved in the engineering of multiple high-performance commercial implementations.  ... 
arXiv:1512.01272v1 fatcat:a4mut4sw4bawjftjbhtignmc2q

Happiness as a Driver of Risk-avoiding Behaviour: Theory and an Empirical Study of Seatbelt Wearing and Automobile Accidents

Robert J. B. Goudie, Sach Mukherjee, Jan-Emmanuel de Neve, Andrew J. Oswald, Stephen Wu
2014 Economica  
Model selection for Bayesian regression models Regression models aim to characterise the relationship between a response variable and a collection of predictor variables.  ...  This idea has a long history and is often attributed to William of Ockham, under the name Occam's razor, or called the principle of parsimony.  ...  This correspondence means that the Metropolis-within-Gibbs move can exploit algorithms designed for variable selection.  ... 
doi:10.1111/ecca.12094 fatcat:65prnkivsnfw7o6ctyzkujheea

An Empirical Bayesian Strategy for Solving the Simultaneous Sparse Approximation Problem

David P. Wipf, Bhaskar D. Rao
2007 IEEE Transactions on Signal Processing  
The resultant algorithm is then compared with multiple response extensions of matching pursuit, basis pursuit, FOCUSS, and Jeffreys prior-based Bayesian methods, finding that it often outperforms the others  ...  Based on the concept of automatic relevance determination, this paper uses an empirical Bayesian prior to estimate a convenient posterior distribution over candidate basis vectors.  ...  Bayesian practitioners have also proposed this idea as a natural means of incorporating the principle of Occam's razor into model selection, often using the description evidence maximization or type-II  ... 
doi:10.1109/tsp.2007.894265 fatcat:osn5pcglmfdvxhlu3xaileaafe

A Framework for Optimization under Limited Information [article]

Tansu Alpcan
2011 arXiv   pre-print
Bayesian approach and using Gaussian processes as a state-of-the-art regression method.  ...  In many real world problems, optimization decisions have to be made with limited information.  ...  Future research directions are abundant and include further investigation of the exploration-exploitation trade-off, adaptive weighting parameters, and random sampling methods for problems in higher dimensional  ... 
arXiv:1105.2176v1 fatcat:lu4jgrlbxbazzbevt4h5wvzuw4

Transforming 21st-century Leader Competencies by Developing Mindfulness

Bogna Gąsiorowska, WSB University in Gdansk, Paweł Gąsiorowski
2019 e-mentor  
We work in partnership with others, including governments, businesses and charities. We are a UK charity that works all over the world, supported by a financial endowment.  ...  Questions for the authors Questions for the authors can be sent to: hasan.bakhshi@nesta.org.uk (Hasan Bakhshi) jd@robots. ox.ac.uk (Jonathan M. Downing) mosb@robots.ox.ac.uk (Michael A.  ...  ACKNOWLEDGEMENTS We thank Logan Graham for his research assistance on coding and preparation of the O*NET and employment data and to Justin Bewsher for his contributions to the sensitivity analysis.  ... 
doi:10.15219/em82.1445 fatcat:r5ulrmjupfdlfcm4hfar5f6bta
« Previous Showing results 1 — 15 out of 82 results