Modeling of dead time systems by support vector regression

Various techniques such as neural networks and fuzzy systems are widely used in modeling of dead time systems frequently encountered in engineering problems. Support vector machine, a new learning method, is a nonlinear kernel based learning machine. It was found that this method causes good results in solving the problems in the areas of modeling, system identification, classification, characterizing writing, regression, function estimation and optimal control. In this study, it was examined that the principles of the theory of regression by SVM, which demonstrated a good generalization skill in the test group in which various inputs also used in training were applied, and modeling problem of dead time systems relating to its usage.