Subset selection in sparse matrices

Alberto Del Pia (delpiawisc.edu)
Santanu Dey (santanu.deyisye.gatech.edu)
Robert Weismantel (robert.weismantelifor.math.ethz.ch)

Abstract: In subset selection we search for the best linear predictor that involves a small subset of variables. From a computational complexity viewpoint, subset selection is NP-hard and few classes are known to be solvable in polynomial time. Using mainly tools from discrete geometry, we show that some sparsity conditions on the original data matrix allow us to solve the problem in polynomial time.

Keywords: subset selection, linear regression, polynomial-time algorithm, sparsity

Category 1: Nonlinear Optimization (Nonlinear Systems and Least-Squares )

Category 2: Global Optimization (Theory )

Citation: Submitted manuscript

Download: [PDF]

Entry Submitted: 10/05/2018
Entry Accepted: 10/05/2018
Entry Last Modified: 02/03/2020

Modify/Update this entry