Randomized maximum-contrast selection: subagging for large-scale regression
From MaRDI portal
Abstract: We introduce a very general method for sparse and large-scale variable selection. The large-scale regression settings is such that both the number of parameters and the number of samples are extremely large. The proposed method is based on careful combination of penalized estimators, each applied to a random projection of the sample space into a low-dimensional space. In one special case that we study in detail, the random projections are divided into non-overlapping blocks; each consisting of only a small portion of the original data. Within each block we select the projection yielding the smallest out-of-sample error. Our random ensemble estimator then aggregates the results according to new maximal-contrast voting scheme to determine the final selected set. Our theoretical results illuminate the effect on performance of increasing the number of non-overlapping blocks. Moreover, we demonstrate that statistical optimality is retained along with the computational speedup. The proposed method achieves minimax rates for approximate recovery over all estimators using the full set of samples. Furthermore, our theoretical results allow the number of subsamples to grow with the subsample size and do not require irrepresentable condition. The estimator is also compared empirically with several other popular high-dimensional estimators via an extensive simulation study, which reveals its excellent finite-sample performance.
Recommendations
Cites work
- A note on methods of restoring consistency to the bootstrap
- Adaptive estimation of a quadratic functional by model selection.
- Analyzing bagging
- Bagging predictors
- Bootstrap methods: another look at the jackknife
- Concentration inequalities for dependent random variables via the martingale method
- Concentration of Lipschitz Functionals of Determinantal and Other Strong Rayleigh Measures
- Divide and conquer kernel ridge regression: a distributed algorithm with minimax optimal rates
- Exchangeable Bernoulli random variables and Bayes postulate
- Exchangeably weighted bootstraps of the general empirical process
- Finite exchangeable sequences
- Generalized Pólya urn Designs with Null Balance
- Geometric median and robust estimation in Banach spaces
- scientific article; zbMATH DE number 1220667 (Why is no real title available?)
- Local case-control sampling: efficient subsampling in imbalanced data sets
- On the asymptotic theory of subsampling
- On the conditions used to prove oracle results for the Lasso
- Oracle inequalities and optimal inference under group sparsity
- Restricted eigenvalue properties for correlated Gaussian designs
- Robust and scalable Bayes via a median of subset posterior measures
- Simultaneous analysis of Lasso and Dantzig selector
- Sup-norm convergence rate and sign concentration property of Lasso and Dantzig estimators
- Tight conditions for consistency of variable selection in the context of high dimensionality
Cited in
(6)- A MOM-based ensemble method for robustness, subsampling and hyperparameter tuning
- Maximin effects in inhomogeneous large-scale data
- Sampling Lasso quantile regression for large-scale data
- Semi-Standard Partial Covariance Variable Selection When Irrepresentable Conditions Fail
- RaSE: A Variable Screening Framework via Random Subspace Ensembles
- The COR criterion for optimal subset selection in distributed estimation
This page was built for publication: Randomized maximum-contrast selection: subagging for large-scale regression
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5965320)
