A branch and bound approach for a class of non-convex problems with applications to robust regression

摘要

We consider a class of non-convex problems, with application to robust regression and robust support vector machines. We propose an algorithm that computes the exact solution using a branch and bound approach in parameter space. Numerical experiments show that, in some cases, the time complexity of the algorithm is linear with respect to the number of samples, while it is exponential with respect to the number of regressors.