A Novel Framework for Online Supervised Learning with Feature Selection
This addresses the problem of efficient online learning with feature selection for large-scale data applications, representing a novel method for a known bottleneck rather than incremental work.
The paper tackles the problem of lower convergence rates and limited feature selection capability in online learning by proposing a novel framework based on running averages that creates online versions of popular offline regularized methods like Lasso and Elastic Net. The result includes proved equivalence to offline counterparts, theoretical guarantees for support recovery and convergence, and numerical experiments showing high true support recovery accuracy and faster convergence rates compared to conventional algorithms.
Current online learning methods suffer issues such as lower convergence rates and limited capability to select important features compared to their offline counterparts. In this paper, a novel framework for online learning based on running averages is proposed. Many popular offline regularized methods such as Lasso, Elastic Net, Minimax Concave Penalty (MCP), and Feature Selection with Annealing (FSA) have their online versions introduced in this framework. The equivalence between the proposed online methods and their offline counterparts is proved, and then novel theoretical true support recovery and convergence guarantees are provided for some of the methods in this framework. Numerical experiments indicate that the proposed methods enjoy high true support recovery accuracy and a faster convergence rate compared with conventional online and offline algorithms. Finally, applications to large datasets are presented, where again the proposed framework shows competitive results compared to popular online and offline algorithms.