This paper considers nonlinear modeling based on a limited amount of experimental data and a simulator built from prior knowledge. The problem of how to best incorporate the data ...
—Adapting the hyperparameters of support vector machines (SVMs) is a challenging model selection problem, especially when flexible kernels are to be adapted and data are scarce....
Background: Elucidating biological networks between proteins appears nowadays as one of the most important challenges in systems biology. Computational approaches to this problem ...
Pierre Geurts, Nizar Touleimat, Marie Dutreix, Flo...
Background: This paper introduces the notion of optimizing different norms in the dual problem of support vector machines with multiple kernels. The selection of norms yields diff...
Many complex, real world phenomena are difficult to study directly using controlled experiments. Instead, the use of computer simulations has become commonplace as a feasible alte...