Abstract: Smooth Support Vector Machine(SSVM),which has better advantage than Support Vector Machine(SVM)，is a model of SVM for ctuick solving. We got a new smooth function for approximating the plus function by interpolation base on smooth piecewise polynomial functions. We found a new model of Smooth Support Vector Machine by improving the model of Polynomial Smooth Support Vector Machine using the new smooth function. Performances of approaching and approximate error were given for the new smooth function, as well as the study of convergence and the approximation limit of optimum for the new model. Obviously numeric experiment shows that the new model has better performance than PSSVM.

Key words: Pattern recognition, Polynomial smooth support vector machine, Spline polynomial, Smooth technology