DC models for spherical separation

@article{Astorino2010DCMF,
  title={DC models for spherical separation},
  author={Annabella Astorino and Antonio Fuduli and Manlio Gaudioso},
  journal={J. Global Optimization},
  year={2010},
  volume={48},
  pages={657-669}
}
We propose two different approaches for spherical separation of two sets. Both methods are based on minimizing appropriate nonconvex nondifferentiable error functions, which can be both expressed in a DC (Difference of two Convex) form. We tackle the problem by adopting the DC-Algorithm. Some numerical results on classical binary datasets are reported.