一种结合Vague集的模糊支持向量机研究

Fuzzy SVM Based on Vague Sets

为了解决传统支持向量机对噪声或野值敏感的问题,模糊支持向量机给出一种解决办法,就是区别对待训练样本,为每一个数据点分配不同的权重,使其在分类模型训练过程中起不同的作用.以期获取更加合理的分类超平面,使得分类模型具有更好的泛化能力.Vague隶属度的计算是该算法实现的关键步骤之一,文中给出一种基于模糊C-均值聚类方法的Vague隶属度计算的方法,可以生成训练样本的真、假隶属度.实验结果表明,该算法具有更好的抗噪性能及分类能力.

In order to classify data with noises or outliers, Fuzzy support vector machine （FSVM） improve the generalization power of traditional SVM by assigning a fuzzy membership to each input data point. In this paper, an improved FSVM based on vague sets is proposed by assigning a truth-membership and a false-membership to each data point. And we reformulate the improved FSVM so that different input points can make different contributions to decision hyperplane. Furthermore, the key part of FSVM based on vague sets is how to construct the vague membership in the training data set. Fuzzy C-means clustering algorithm is employed for generating the truth-membership and false-membership. The experimental result shows the proposed model is promising.