节点文献
一种求支持向量机光滑函数的新方法
New Method for Solving Smoothing Functions of Support Vector Machines
【摘要】 光滑函数在支持向量机中起着重要作用。用插值函数的方法曾导出了一个求光滑函数的递推公式,然而,用该递推公式求光滑函数还很繁琐。针对该问题,用积分的方法得到了求多项式光滑函数的一个新递推公式。算例表明,用新递推公式求光滑函数比原递推公式简单得多,为研究支持向量机的光滑函数提供了一种新方法。
【Abstract】 Smoothing functions play an important role in Support Vector Machines(SVMs).A recursive formula with the technique of interpolation functions was developed to get smoothing polynomial functions.It is still very tedious to get smooth functions using the recursive formula.Aiming at this problem,a new recursive formula of smoothing functions was got using the method of integral.The computing examples show that it is much simpler to get smoothing functions with new recursive formula than with old one.An important property of the smooth polynomial functions was deduced with the new recursive formula,which provided a new method for studying smoothing functions of SVM.
【Key words】 pattern recognition; support vector machine; smoothing function; classification;
- 【文献出处】 系统仿真学报 ,Journal of System Simulation , 编辑部邮箱 ,2008年15期
- 【分类号】TP301.6
- 【被引频次】12
- 【下载频次】143