Name: David B. Dunson
Type: User
Company: Department of Statistical Science, Duke University
Bio: Develops statistical & machine learning methods for complex data, with a particular focus on Bayesian approaches and applications in ecology & biomedicine
Twitter: DavidDunson7
Location: Box 90251, Duke University, Durham, NC 27708
Blog: http://www2.stat.duke.edu/~dunson/
David B. Dunson's Projects
This project provides codes for manifold learning using spherelets (see https://arxiv.org/abs/1706.08263 )
"Adverse Subpopulation Regression for Multivariate Outcomes with High-Dimensional Predictors" (2011) by Bin Zhu, David B. Dunson, Allison E. Ashley-Koch
Contains code from the paper: Zito A, Rigon T, Dunson DB (2022) "Inferring taxonomic placement from DNA barcoding allowing discovery of new species" arXiv:2201.09782. DNA barcoding is conducted on field samples and it is important to classify the samples taxonomically, while allowing discovery of new taxa; these may be organisms unknown to science or ones known to science by lacking a reference sequence. BayesANT allows the current taxonomy from the reference database to grow probabilistically as new DNA barcoding data are collected. This work was motivated by our Lifeplan project funded by the ERC.
"Bayesian Consensus Clustering" (2013) by Eric F. Lock and David B. Dunson.
"Bayesian Constraint Relaxation" by Duan, Leo L., Alexander L. Young, Akihiko Nishimura, and David B. Dunson
Bayesian Conditional Tensor Factorization
This paper proposes a new class of Bayesian species sampling models motivated by DNA barcoding data. The fundamental problem is predicting how many new OTUs will be present in some additional number of sequences based on partial sequencing information. All the biological sample can't be sequenced due to expense issues so this allows inference on how many OTUs are in the sample based on limited sequencing depth.
Matlab code for the paper Gu, Y. and Dunson, D.B. (2021), Identifying Interpretable Discrete Latent Structures from Discrete Data.
"Shared kernel Bayesian screening" (2015) & "Bayesian genome- and epigenome-wide association studies with gene level dependence" (2016) by Lock, E. F. and Dunson, D. B.
Bacth corrected t-SNE - Aliverti, Wilhelmsen and Dunson
Bayesian Modular and Multiscale Regression
Bayesian Nonparametric Modeling of Higher Order Markov Chains
"Robust Bayesian inference via coarsening" (2015) by J. W. Miller and D. B. Dunson.
"Framework for Probabilistic Inferences from Imperfect Models" by Meng Li, David B. Dunson
'Dirichlet Laplace prior for optimal shrinkage" by Bhattacharya, Pati, Pillai and Dunson
"Discontinuous Hamiltonian Monte Carlo for sampling discrete parameters" by Akihiko Nishimura, David Dunson, Jianfeng Lu
WASP and PIE algorithms of Sanvesh Srivastava, Cheng Li, and David Dunson (2015, 2017)
Density estimation using Fisher-Gaussian (FG) kernel mixture approach
"Bayesian Gaussian Copula Factor Models for Mixed Data" by Jared S. Murray, David B. Dunson, Lawrence Carin, Joseph E. Lucas. This is a read-only mirror of the CRAN R package repository.
This contains an R package for implementing the methods in Zhang R, Mak S, Dunson DB (2022) Gaussian process subspace prediction for model reduction. SIAM Journal on Scientific Computing 44 (3), A1428-A1449
This repository contains code to implement the methods from the paper Schiavon, Canale and Dunson (2022), "Generalized infinite factorization models", Biometrika 109 (3), 817-835. This article proposes a novel class of structured Bayesian latent factor models which allow one to include "meta features" providing information on the different measured variables; such features can inform about the dependence structure among the variables.
"Generalized Admixture Mapping for Complex Traits" (2011) by Bin Zhu, Allison E. Ashley-Koch, David B. Dunson
Bayesian infinite factor modelling
This repository contains code from the paper Chakraborty A, Ovaskainen O, Dunson DB (2022) "Bayesian semi parametric long memory models for discretized event data", Annals of Applied Statistics 16 (3), 1380-1399. The paper proposes a novel fractional Brownian probit model for binary time series with long memory motivated by ecological applications to bird species monitoring.
"Bayesian Local Extrema Splines" by Matthew W. Wheeler, David B. Dunson, Amy H. Herring.
Bayesian Local Extremum Splines
"Fast moment estimation for generalized latent Dirichlet models" (2016) by Shiwen Zhao, Barbara E. Engelhardt, Sayan Mukherjee, David B. Dunson