Parallel decomposition approaches for training support vector machines
Thomas Serafini, Gaetano Zanghirati and Luca Zanni
Parallel Computing: Software Technology, Algorithms, Architectures and Applications
Advances in Parallel Computing
, Amsterdam, The Netherlands
We consider parallel decomposition techniques for solving the large quadratic programming (QP) problems arising in training support vector machines. A recent technique is improved by introducing an efficient solver for the inner QP subproblems and a preprocessing step useful to hot start the decomposition strategy. The effectiveness of the proposed improvements is evaluated by solving large-scale benchmark problems on different parallel architectures.