Filters








1,806 Hits in 2.0 sec

Detecting Hierarchical Changes in Latent Variable Models [article]

Shintaro Fukushima, Kenji Yamanishi
2020 arXiv   pre-print
The key idea to realize it is to employ the MDL (minimum description length) change statistics for measuring the degree of change, in combination with DNML (decomposed normalized maximum likelihood) code-length  ...  This paper addresses the issue of detecting hierarchical changes in latent variable models (HCDL) from data streams.  ...  DNML Code-Length In order to resolve the problem in the previous section, we employ the decomposed normalized maximum likelihood (DNML) code-length for latent variable models instead of the NML code-length  ... 
arXiv:2011.09465v3 fatcat:c3ztgpckfja4dkg7r45ysb6dbi

BayesSUR: An R package for high-dimensional multivariate Bayesian variable and covariance selection in linear regression [article]

Zhi Zhao, Marco Banterle, Leonardo Bottolo, Sylvia Richardson, Alex Lewin, Manuela Zucknick
2021 arXiv   pre-print
Several variable selection priors have been implemented in this context, in particular the hotspot detection prior for latent variable inclusion indicators, which results in sparse variable selection for  ...  The R package allows the specification of the models in a modular way, where the user chooses the priors for variable selection and for covariance selection separately.  ...  BayesSUR Health" (AL), UK Medical Research Council grants MR/M013138/1 (MB, AL, LB, SR) and MC_UU_00002/10 (SR), NIHR Cambridge BRC (SR), BHF-Turing Cardiovascular Data Science Awards 2017 (LB) and The  ... 
arXiv:2104.14008v1 fatcat:vzz47sgmg5gq7kr3jhjjzzriw4

Simulation Study on Local Influence Diagnosis for Poisson Mixed-Effect Linear Model [article]

N.Zhang
2018 arXiv   pre-print
Given that hierarchical count data in many fields are not Normally-distributed and include random effects, this paper extends the Generalized Linear Mixed Models (GLMMs) into Poisson Mixed-Effect Linear  ...  Model (PMELM) and do numerical simulation experiments to verify the approach proposed by Rakhmawati et al. (2016) in detecting outliers.  ...  Rakmawati et al.(2016)applies it in GLMM, the local influence for the Probit-normal Model could be interpreted by fixed effect and squared of variability of random effect.  ... 
arXiv:1805.02898v1 fatcat:fmvkg3i3hbhfpl4tkfaaq3n2vu

Auto-Encoding Total Correlation Explanation [article]

Shuyang Gao, Rob Brekelmans, Greg Ver Steeg, Aram Galstyan
2018 arXiv   pre-print
This information-theoretic view of VAE deepens our understanding of hierarchical VAE and motivates a new algorithm, AnchorVAE, that makes latent codes more interpretable through information maximization  ...  Advances in unsupervised learning enable reconstruction and generation of samples from complex distributions, but this success is marred by the inscrutability of the representations learned.  ...  likelihood of x under this model.  ... 
arXiv:1802.05822v1 fatcat:3jbmk6h2lndjhi6mypmy2lhqy4

A Survey of Model Evaluation Approaches With a Tutorial on Hierarchical Bayesian Methods

Richard Shiffrin, Michael Lee, Woojae Kim, Eric-Jan Wagenmakers
2008 Cognitive Science  
This article reviews current methods for evaluating models in the cognitive sciences, including theoretically based approaches, such as Bayes factors and minimum description length measures; simulation  ...  This article argues that hierarchical methods, generally, and hierarchical Bayesian methods, specifically, can provide a more thorough evaluation of models in the cognitive sciences.  ...  The minimal code length used is the combined code length for the model and data described by the model. Various MDL measures have been developed as approximations to this code length.  ... 
doi:10.1080/03640210802414826 pmid:21585453 fatcat:tu7ie3lmjzblboun7mrml7575q

A Guide toCoCo

Jens Henrik Badsberg
2001 Journal of Statistical Software  
CoCo finds closed form expres-iv sion for estimates in decomposable models, handles incomplete tables and tables with incomplete observations, computes exact tests between any two nested decomposable models  ...  DIGRAM handles recursive graphical models on contingency tables. MIM also handles continuous variables in mixed interaction models, CG-distributions.  ...  Letp(i A ) be the maximum likelihood estimate on the table (n(i A )) i A ∈IA for a model M A , where the union A of variables in the model is a subset of the declared variables ∆.p(i) is the maximum likelihood  ... 
doi:10.18637/jss.v006.i04 fatcat:ukq7o2t3vreb3djzuvrcu5oade

Heteroscedastic factor mixture analysis

Angela Montanari, Cinzia Viroli
2010 Statistical Modelling  
In this paper the model is presented and a maximum likelihood estimation procedure via the EM algorithm is developed.  ...  By replacing the traditional assumption of Gaussian distributed factors by a finite mixture of multivariate Gaussians, the unobserved heterogeneity can be modelled by latent classes.  ...  In the next section maximum likelihood estimation for the HFMA model through an EM algorithm is developed and illustrated.  ... 
doi:10.1177/1471082x0901000405 fatcat:azpgeou64nhlnowq5avrsdku6m

Hierarchical clustering with discrete latent variable models and the integrated classification likelihood [article]

Etienne Côme, Nicolas Jouvin, Pierre Latouche, Charles Bouveyron
2020 arXiv   pre-print
Considering the integrated classification likelihood criterion as an objective function, this work applies to every discrete latent variable models (DLVMs) where this quantity is tractable.  ...  In this paper, we introduce a general two-step methodology for model-based hierarchical clustering.  ...  Conclusion In this paper, we proposed a new methodology for model-based hierarchical clustering with discrete latent variables models, based on two related contributions.  ... 
arXiv:2002.11577v2 fatcat:raqrkrfz3bevtob7bwuntjsyvq

Non-parametric frailty Cox models for hierarchical time-to-event data

Francesca Gasperoni, Francesca Ieva, Anna Maria Paganoni, Christopher H Jackson, Linda Sharples
2018 Biostatistics  
A tailored Expectation-Maximization algorithm is proposed for estimating the model parameters, methods of model selection are compared, and the code is assessed in simulation studies.  ...  We propose a novel model for hierarchical time-to-event data, for example, healthcare data in which patients are grouped by their healthcare provider.  ...  Our model identified three latent populations using the BIC for selection, four with AIC and six with Laird's criterion (Laird, 1978) .  ... 
doi:10.1093/biostatistics/kxy071 pmid:30590499 pmcid:PMC6451633 fatcat:3jpr3i5ugvaovdttm2isju6ama

Dimensionality Reduction of Clustered Data Sets

Guido Sanguinetti
2008 IEEE Transactions on Pattern Analysis and Machine Intelligence  
We prove that the maximum likelihood solution of the model is an unsupervised generalization of linear discriminant analysis.  ...  We present a novel probabilistic latent variable model to perform linear dimensionality reduction on data sets which contain clusters.  ...  ACKNOWLEDGMENTS The author would like to thank Mahesan Niranjan and Mark Girolami for useful discussions and suggestions.  ... 
doi:10.1109/tpami.2007.70819 pmid:18195446 fatcat:5it4oacla5fptket7mg2ipkrgm

A Generative Model for Measuring Latent Timing Structure in Motor Sequences

Christopher M. Glaze, Todd W. Troyer, Melissa J. Coleman
2012 PLoS ONE  
In addition to quantifying the variability contributed by each of these latent factors in the data, the approach assigns maximum likelihood estimates of each factor on a trial-to-trial basis.  ...  Across our sample of syllables, both global and independent variability scale with average length while timing jitter does not, a pattern consistent with the Wing and Kristofferson (1973) model of sequence  ...  Christina Castellino for helpful comments. Author Contributions Conceived and designed the experiments: CMG TWT. Performed the experiments: CMG TWT. Analyzed the data: CMG TWT.  ... 
doi:10.1371/journal.pone.0037616 pmid:22815683 pmcid:PMC3398023 fatcat:voldl6epkrbetmo3hcazstyfeu

Feature-Based Pose Estimation [chapter]

Cristian Sminchisescu, Liefeng Bo, Catalin Ionescu, Atul Kanaujia
2011 Visual Analysis of Humans  
In this context, we discuss several predictive methods including large-scale mixture of experts, supervised spectral latent variable models and structural support vector machines, asses the impact of the  ...  We argue that reliable 3d human pose prediction can be achieved through an alliance between image descriptors that encode multiple levels of selectivity and invariance and models that are capable to represent  ...  Acknowledgments: This work has been supported in part by the European Commission under MCEXT-025481 and by CNCSIS-UEFISCU under project PN II-RU-RC-2/2009.  ... 
doi:10.1007/978-0-85729-997-0_12 fatcat:6ssbr5xz3renfa735skoyw5rxi

Editors' introduction

1983 Journal of Econometrics  
They discuss the elegant maximum likelihood theory for hierarchical loglinear models, and the equally elegant iterative proportional fitting procedure.  ...  We code the categorical variables by using dummies, i.e., binary vectors of length kj in which exactly one element is equal to one.  ... 
doi:10.1016/0304-4076(83)90091-x fatcat:675y3juydbgolcj7iyedeqcvom

Variational Predictive Routing with Nested Subjective Timescales [article]

Alexey Zakharov, Qinghai Guo, Zafeirios Fountas
2021 arXiv   pre-print
change, thus modeling continuous data as a hierarchical renewal process.  ...  By employing an event detection mechanism that relies solely on the system's latent representations (without the need of a separate model), VPR is able to dynamically adjust its internal state following  ...  ACKNOWLEDGMENTS The authors would like to thank Yansong Chua for his valuable contributions and comments made for the earlier versions of the model presented in this manuscript.  ... 
arXiv:2110.11236v1 fatcat:g5ian2kx6fhspfvlvossp3pb5q

Model order selection for boolean matrix factorization

Pauli Miettinen, Jilles Vreeken
2011 Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining - KDD '11  
The technique has desirable properties, such as high interpretability and natural sparsity. But so far no method for selecting the correct model order for BMF has been available.  ...  We formulate the description length function for BMF in generalmaking it applicable for any BMF algorithm.  ...  MDL FOR BMF In this section we give our approach for selecting model orders for BMF by the Minimum Description Length principle.  ... 
doi:10.1145/2020408.2020424 dblp:conf/kdd/MiettinenV11 fatcat:pb7xvqypa5ayff5gvwis4uc2wu
« Previous Showing results 1 — 15 out of 1,806 results