A comparison of four outlier-robust regression formulations using mixed integer optimization in Julia
This code was written by Ali Borenstein and Austin Zaccor as part of a final project in Machine Learning Under a Modern Optimization Lens at MIT as part of their Masters of Business Analytics degree. Please reach out to either of us on github or LinkedIn with questions about the code or mathematical formulations. As mentioned in the main file, publication is forthcoming.
Synthetic datasets are available in Data folder, synthetic data generation process available in "Data Generation Function Final.ipynb"
Other datasets used in this file are as follows:
Hertzsprung-Russell Start Dataset http://support.sas.com/documentation/cdl/en/imlug/64248/HTML/default/viewer.htm#imlug_robustregexpls_sect003.htm
Stackloss of Brownlee: http://support.sas.com/documentation/cdl/en/imlug/65547/HTML/default/viewer.htm#imlug_robustregexpls_sect012.htm
QSAR Aquatic Toxicity (Retrieved from UCI Machine Learning Repository): https://archive.ics.uci.edu/ml/datasets/QSAR+aquatic+toxicity