Tuning Support Vector Machines and Boosted Trees Using Optimization Algorithms

Statistical learning methods have been growing in popularity recent years. Many of these procedures parameters that must be tuned for models to perform well. Research has extensive neural networks, but not many other methods. We looked at the behavior tuning support vector machines, gradient boosting and adaboost both a classification regression setting. used grid search identify ranges where good can found across different datasets. then explored optimization algorithms select model parameter space. Models selected by algorithm were compared best obtained through well performing algorithms. This information was create an R package, EZtune, automatically tunes machines boosted trees.