Polynomial smoothing support vector regression
Lianglun Cheng · Control theory & applications · 2011
Smoothing functions can transform the unsmooth support vector regressions into smooth ones,and thus better regression results can be obtained.It has been one of the key problems to seek a better smoothing function in this field for a long time.Using the series expansion,a new class of polynomial smoothing functions is proposed for the |x|e2 function in e-insensitive support vector regressions.Their important properties are then discussed.It is shown that the approximation accuracy and smoothness rank of polynomial functions can be as high as required.The experimental results show that as the smoothness rank of polynomial functions increases,the approximation accuracy and the regression performance are correspondingly improved.Therefore,the new class of polynomial functions provides better performance for smoothing the support vector regression.