| Paper: | SPTM-P13.6 | ||
| Session: | Detection, Estimation, Classification Theory and Applications | ||
| Time: | Friday, May 19, 14:00 - 16:00 | ||
| Presentation: | Poster | ||
| Topic: | Signal Processing Theory and Methods: Detection, Estimation, Classification Theory and Applications | ||
| Title: | AN INFO-GAP APPROACH TO LINEAR REGRESSION | ||
| Authors: | Miriam Zacksenhouse, Simona Nemets, Anna Yoffe, Yakov Ben-Haim, Technion - Israel Institute of Technology, Israel; Mikhail Lebedev, Miguel Nicolelis, Duke University, United States | ||
| Abstract: | Linear regression with high uncertainties is a major challenging problem. Standard regression techniques are based on optimizing performance and are highly sensitive to uncertainties. Regularization techniques depend on proper parameter selection, which is also problematic under severe uncertainties. Here we develop an alternative regression methodology based on satisficing rather than optimizing the performance criterion while maximizing the robustness to uncertainties. Uncertainties are represented by info-gap models which entail an unbounded family of nested sets of measurements parameterized by a non-probabilistic horizon of uncertainty. We prove and demonstrate that the robust-satisficing solution is different from the optimal least squares solution and that the info-gap approach can provide higher robustness to uncertainty. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012