Multi-stage Convex Relaxation for Feature Selection

Abstract

A number of recent work studied the effectiveness of feature selection using Lasso. It is known that under the restricted isometry properties (RIP), Lasso does not generally lead to the exact recovery of the set of nonzero coefficients, due to the looseness of convex relaxation. This paper considers the feature selection property of nonconvex regularization, where the solution is given by a multi-stage convex relaxation scheme. Under appropriate conditions, we show that the local solution obtained by this procedure recovers the set of nonzero coefficients without suffering from the bias of Lasso relaxation, which complements parameter estimation results of this procedure.

References in corpus (1)

• Nearly unbiased variable selection under minimax concave penalty

Cited by in corpus (5)

• Multi-Stage Multi-Task Feature Learning
• Efficient Sparse Group Feature Selection via Nonconvex Optimization
• Combined l_1 and greedy l_0 penalized least squares for linear model selection
• Confidence Intervals for Low-Dimensional Parameters in High-Dimensional Linear Models
• A General Theory of Concave Regularization for High Dimensional Sparse Estimation Problems