Current browse context:
cs.LG
Change to browse by:
References & Citations
Statistics > Computation
Title: Local Uncertainty Sampling for Large-Scale Multi-Class Logistic Regression
(Submitted on 27 Apr 2016 (v1), last revised 13 Sep 2018 (this version, v3))
Abstract: A major challenge for building statistical models in the big data era is that the available data volume far exceeds the computational capability. A common approach for solving this problem is to employ a subsampled dataset that can be handled by available computational resources. In this paper, we propose a general subsampling scheme for large-scale multi-class logistic regression and examine the variance of the resulting estimator. We show that asymptotically, the proposed method always achieves a smaller variance than that of the uniform random sampling. Moreover, when the classes are conditionally imbalanced, significant improvement over uniform sampling can be achieved. Empirical performance of the proposed method is compared to other methods on both simulated and real-world datasets, and these results match and confirm our theoretical analysis.
Submission history
From: Lei Han [view email][v1] Wed, 27 Apr 2016 15:00:12 GMT (289kb,D)
[v2] Thu, 16 Jun 2016 09:05:18 GMT (598kb,D)
[v3] Thu, 13 Sep 2018 05:08:08 GMT (1454kb,D)
Link back to: arXiv, form interface, contact.