一种新颖隶属度函数的模糊支持向量机

A Novel Membership Function for Fuzzy Support Vector Machines

传统的支持向量机(SVM)训练含有外部点或噪音数据时,容易产生过拟合(over-fitting)。通过模糊隶属度函数来降低外部点或被污染数据的选择。本文提出了一种新的核隶属度函数,这种新的隶属度函数不仅依赖于每个样本点到类型中心的距离,还依赖于该样本点最邻近的K个其他样本点的距离。实验结果表明了具有该隶属度函数的模糊支持向量机的有效性。

The Traditional Support Vector Machine often falls into over-fitting when outliers are contained in the training data. The Fuzzy Support Vector Machine can effectively deal with the over-fitting problem by reducing the effect of the outliers using the fuzzy membership function. The performance of the fuzzy support vector machine is largely dependent on the choice of fuzzy membership functions. In this paper, we propose a new kernel membership function, which is not only dependent on the distance between each data and the center of the class, but also related to its K nearest data. Experiments validate the good performance of our novel kernel membership function.