1 paper
Martin Huang, Samuel Muller, Garth Tarr
Best subset selection procedures typically rely on a squared error loss, where a small number of outlying observations may distort the entire solution path. Replacing this loss wit…