Robust and efficient subsampling algorithms for massive data logistic regression
- 26 April 2023
- journal article
- research article
- Published by Taylor & Francis in Journal of Applied Statistics
- Vol. 51 (8) , 1427-1445
Abstract
Datasets that are big with regard to their volume, variety and velocity are becoming increasingly common. However, limitations in computer processing can often restrict analysis performed on them. Nonuniform subsampling methods are effective in reducing computational loads for massive data. However, the variance of the estimator of nonuniform subsampling methods becomes large when the subsampling probabilities are highly heterogenous. To this end, we develop two new algorithms to improve the estimation method for massive data logistic regression based on a chosen hard threshold value and combining subsamples, respectively. The basic idea of the hard threshold method is to carefully select a threshold value and then replace subsampling probabilities lower than the threshold value with the chosen value itself. The main idea behind the combining subsamples method is to better exploit information in the data without hitting the computation bottleneck by generating many subsamples and then combining estimates constructed from the subsamples. The combining subsamples method obtains the standard error of the parameter estimator without estimating the sandwich matrix, which provides convenience for statistical inference in massive data, and can significantly improve the estimation efficiency. Asymptotic properties of the resultant estimators are established. Simulations and analysis of real data are conducted to assess and showcase the practical performance of the proposed methods.Keywords
Funding Information
- National Natural Science Foundation of China (#11471264)
This publication has 21 references indexed in Scilit:
- Randomized sketches for kernels: Fast and optimal nonparametric regressionThe Annals of Statistics, 2017
- Local case-control sampling: Efficient subsampling in imbalanced data setsThe Annals of Statistics, 2014
- A Scalable Bootstrap for Massive DataJournal of the Royal Statistical Society Series B: Statistical Methodology, 2014
- Low rank approximation and regression in input sparsity time2013
- Using balanced iterative reducing and clustering hierarchies to compute approximate rank statistics on massive datasetsJournal of Statistical Computation and Simulation, 2013
- Robust learning from bites for data miningComputational Statistics & Data Analysis, 2007
- Tuning parameter selectors for the smoothly clipped absolute deviation methodBiometrika, 2007
- Sampling algorithms for l2 regression and applications2006
- Target estimation for the logistic regression modelJournal of Statistical Computation and Simulation, 2005
- Experiments with mixtures: Optimal allocations for becker’s modelsMetrika, 1997