Stochastic Discrete First-Order Algorithm for Feature Subset Selection

Kota KUDO  Yuichi TAKANO  Ryo NOMURA  

IEICE TRANSACTIONS on Information and Systems   Vol.E103-D   No.7   pp.1693-1702
Publication Date: 2020/07/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.2019EDP7274
Type of Manuscript: PAPER
Category: Artificial Intelligence, Data Mining
feature subset selection,  optimization algorithm,  linear regression,  machine learning,  statistics,  

Full Text: PDF(754.8KB)>>
Buy this Article

This paper addresses the problem of selecting a significant subset of candidate features to use for multiple linear regression. Bertsimas et al. [5] recently proposed the discrete first-order (DFO) algorithm to efficiently find near-optimal solutions to this problem. However, this algorithm is unable to escape from locally optimal solutions. To resolve this, we propose a stochastic discrete first-order (SDFO) algorithm for feature subset selection. In this algorithm, random perturbations are added to a sequence of candidate solutions as a means to escape from locally optimal solutions, which broadens the range of discoverable solutions. Moreover, we derive the optimal step size in the gradient-descent direction to accelerate convergence of the algorithm. We also make effective use of the L2-regularization term to improve the predictive performance of a resultant subset regression model. The simulation results demonstrate that our algorithm substantially outperforms the original DFO algorithm. Our algorithm was superior in predictive performance to lasso and forward stepwise selection as well.