Simultaneous Variable Selection and Sensor Selection using Convex Penalties

by Ernest Fokoue.

Abstract: This paper adapts results and tools from the traditional D-optimality machinery to derive an efficient technique for simultaneous variable selection and sensor selection. The proposed approach provides the advantage that the underlying variable selection problem is formulated as a convex optimization task that yields sparse parameter estimates, thereby helping isolate relevant variables efficiently. The theoretical derivation of the proposed method is presented, along with a discussion of its unique benefits.

Key Words: Variable Selection, Convex optimization, D-Optimality, Bayesian analysis, Sensor selection, Sparsity

Author:
Ernest Fokoue, ernest.fokoue@rit.edu

Editor: Khattree, Ravi, khattree@oakland.edu

READING THE ARTICLE: You can read the article in portable document (.pdf) format (83460 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 809 times since APRIL 11, 2013.


Return to the InterStat Home Page.