Back
Sparse regression via a trust-region proximal method
We present a method for sparse regression problems. Our method is based on the nonsmooth trust-region framework that minimizes a sum of smooth convex functions and a nonsmooth convex regularizer. By employing a separable quadratic approximation to the smooth part, the method enables the use of proximity operators, which in turn allow tackling the nonsmooth part efficiently. We illustrate our method by implementing it for three important sparse regression problems. In experiments with synthetic and real-world large-scale data, our method is seen to be competitive, robust, and scalable.
@poster{6522, title = {Sparse regression via a trust-region proximal method}, journal = {24th European Conference on Operational Research (EURO 2010)}, abstract = {We present a method for sparse regression problems. Our method is based on the nonsmooth trust-region framework that minimizes a sum of smooth convex functions and a nonsmooth convex regularizer. By employing a separable quadratic approximation to the smooth part, the method enables the use of proximity operators, which in turn allow tackling the nonsmooth part efficiently. We illustrate our method by implementing it for three important sparse regression problems. In experiments with synthetic and real-world large-scale data, our method is seen to be competitive, robust, and scalable.}, volume = {24}, pages = {278}, organization = {Max-Planck-Gesellschaft}, school = {Biologische Kybernetik}, month = apr, year = {2010}, slug = {6522}, author = {Kim, D. and Sra, S. and Dhillon, I.}, month_numeric = {4} }