Efficient posterior sampling for high-dimensional imbalanced logistic regression

DOI10.1093/BIOMET/ASAA035MaRDI QIDQ5145712zbMATH OpenOpenAlexFDO

Authors Deborshee Sen, Matthias Sachs, Jianfeng Lu, David Dunson

Publication date 21 January 2021

Published in Biometrika (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1905.11232

zbMATH Keywords

imbalanced datalogistic regressionpiecewise-deterministic Markov processscalable inferencesubsampling

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Generalized linear models (logistic models) (62J12) Markov processes: estimation; hidden Markov models (62M05) Applications of statistics to biology and medical sciences; meta analysis (62P10)

Abstract: High-dimensional data are routinely collected in many areas. We are particularly interested in Bayesian classification models in which one or more variables are imbalanced. Current Markov chain Monte Carlo algorithms for posterior computation are inefficient as $n$ and/or $p$ increase due to worsening time per step and mixing rates. One strategy is to use a gradient-based sampler to improve mixing while using data sub-samples to reduce per-step computational complexity. However, usual sub-sampling breaks down when applied to imbalanced data. Instead, we generalize piece-wise deterministic Markov chain Monte Carlo algorithms to include importance-weighted and mini-batch sub-sampling. These approaches maintain the correct stationary distribution with arbitrarily small sub-samples, and substantially outperform current competitors. We provide theoretical support and illustrate gains in simulated and real data applications.

Recommendations

• Local case-control sampling: efficient subsampling in imbalanced data sets
• High dimensional classifiers in the imbalanced case
• MCMC for imbalanced categorical data
• Using synthetic data and dimensionality reduction in high-dimensional classification via logistic regression
• New hard-thresholding rules based on data splitting in high-dimensional imbalanced classification

Cited in

(15)

• High dimensional classifiers in the imbalanced case
• On explicit \(L^2\)-convergence rate estimate for piecewise deterministic Markov processes in MCMC algorithms
• Local uncertainty sampling for large-scale multiclass logistic regression
• Estimating promotion effects in email marketing using a large-scale cross-classified Bayesian joint model for nested imbalanced data
• scientific article; zbMATH DE number 5957445 (Why is no real title available?)
• Improving logistic regression on the imbalanced data by a novel penalized log-likelihood function
• Using synthetic data and dimensionality reduction in high-dimensional classification via logistic regression
• MCMC for imbalanced categorical data
• Stochastic gradient Markov chain Monte Carlo
• Concave-Convex PDMP-based Sampling
• Sampling bias and class imbalance in maximum-likelihood logistic regression
• Emerging directions in Bayesian computation
• Posterior computation with the Gibbs zig-zag sampler
• Ultimate Pólya Gamma Samplers–Efficient MCMC for Possibly Imbalanced Binary and Categorical Data
• Targeted stochastic gradient MCMC for HMMs with rare latent states

This page was built for publication: Efficient posterior sampling for high-dimensional imbalanced logistic regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5145712)