Title: Robust least angle regression and LASSO using bootstrap aggregation (real case study: Iranian poultry farms)

Authors: Reza Behmanesh; Abbas Saghaei

Addresses: Department of Industrial Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran ' Department of Industrial Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran

Abstract: Over the past ten years, many researchers and practitioners have concentrated on the topic of least angle regression, LASSO and forward stagewise (LARS) considerably instead of other subset selection models such as forward and backward stepwise. This model selection algorithm is useful and less greedy in comparison with traditional approaches similar forward stepwise. In this study, we present a taxonomy of all researches that pertains to LARS, and a review of this literature including extensions of it. Next, we introduce a robust LARS by applying bootstrap aggregation to reduce the variance of LARS model. For this purpose, we employ real world case of Iranian poultry farms so as to illustrate our proposed models. The results indicate the proposed hybrid model predicts more accurate compared to classic LARS algorithm. At last, we suggest opportunities for future research in this area.

Keywords: least angle regression; LAR; least absolute shrinkage and selection operator; LASSO; forward stagewise; bootstrap aggregation.

DOI: 10.1504/IJISE.2017.083673

International Journal of Industrial and Systems Engineering, 2017 Vol.26 No.2, pp.201 - 227

Received: 04 Feb 2015
Accepted: 23 Mar 2015
Published online: 19 Apr 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article