Local case-control sampling: efficient subsampling in imbalanced data sets

DOI10.1214/14-AOS1220MaRDI QIDQ480957zbMATH OpenOpenAlexWikidataFDO

Authors William Fithian, Trevor Hastie

Publication date 12 December 2014

Published in The Annals of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1306.3706

zbMATH Keywords

logistic regressionsubsamplingcase-control sampling

Mathematics Subject Classification ID

Point estimation (62F10) Sampling theory, sample surveys (62D05)

Abstract: For classification problems with significant class imbalance, subsampling can reduce computational costs at the price of inflated variance in estimating model parameters. We propose a method for subsampling efficiently for logistic regression by adjusting the class balance locally in feature space via an accept-reject scheme. Our method generalizes standard case-control sampling, using a pilot estimate to preferentially select examples whose responses are conditionally rare given their features. The biased subsampling is corrected by a post-hoc analytic adjustment to the parameters. The method is simple and requires one parallelizable scan over the full data set. Standard case-control sampling is inconsistent under model misspecification for the population risk-minimizing coefficients $het{a}^{\ast }$. By contrast, our estimator is consistent for $het{a}^{\ast }$ provided that the pilot estimate is. Moreover, under correct specification and with a consistent, independent pilot estimate, our estimator has exactly twice the asymptotic variance of the full-sample MLE - even if the selected subsample comprises a miniscule fraction of the full data set, as happens when the original data are severely imbalanced. The factor of two improves to $1+frac1c$ if we multiply the baseline acceptance probabilities by $c>1$ (and weight points with acceptance probability greater than 1), taking roughly $frac1+c2$ times as many data points into the subsample. Experiments on simulated and real data show that our method can substantially outperform standard case-control subsampling.

Recommendations

• Local uncertainty sampling for large-scale multiclass logistic regression
• Surprise sampling: improving and extending the local case-control sampling
• Optimal subsampling for large sample logistic regression
• Optimal subsampling for softmax regression
• More efficient estimation for logistic regression with optimal subsamples

Cited in

(32)

• Conditional characteristic feature screening for massive imbalanced data
• Matrix sketching for supervised classification with imbalanced classes
• Optimal subsampling for softmax regression
• Subdata selection algorithm for linear model discrimination
• Deterministic subsampling for logistic regression with massive data
• Optimal subsampling for large-scale quantile regression
• Semi-supervised inference for case-control binary data under possibly mis-specified logistic models
• Local uncertainty sampling for large-scale multiclass logistic regression
• Post-selection Inference of High-dimensional Logistic Regression Under Case–Control Design
• Fused variable screening for massive imbalanced data
• Optimal subsampling for large‐sample quantile regression with massive data
• Randomized maximum-contrast selection: subagging for large-scale regression
• Optimal subsampling for large sample logistic regression
• More efficient estimation for logistic regression with optimal subsamples
• Optimal Poisson subsampling for softmax regression
• A distance metric-based space-filling subsampling method for nonparametric models
• A two-stage optimal subsampling estimation for missing data problems with large-scale data
• Efficient posterior sampling for high-dimensional imbalanced logistic regression
• Surface temperature monitoring in liver procurement via functional variance change-point analysis
• Surprise sampling: improving and extending the local case-control sampling
• Semi-supervised inference for nonparametric logistic regression
• Likelihood Inference for Large Scale Stochastic Blockmodels With Covariates Based on a Divide-and-Conquer Parallelizable Algorithm With Communication
• A Subsampling Method for Regression Problems Based on Minimum Energy Criterion
• Subsampling in longitudinal models
• Unweighted estimation based on optimal sample under measurement constraints
• Efficient modelling of presence-only species data via local background sampling
• Model constraints independent optimal subsampling probabilities for softmax regression
• Estimating promotion effects in email marketing using a large-scale cross-classified Bayesian joint model for nested imbalanced data
• Rather “Good In, Good Out” Than “Garbage In, Garbage Out”: A Comparison of Various Discrete Subsampling Algorithms Using COVID-19 Data Without a Response Variable
• A semiparametric method for risk prediction using integrated electronic health record data
• A two-part measurement error model to estimate participation in undeclared work and related earnings
• A review on design inspired subsampling for big data

This page was built for publication: Local case-control sampling: efficient subsampling in imbalanced data sets

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q480957)