基于高斯分布的非平衡FSVM

针对传统模糊支持向量机算法采用样本到类中心的距离关系来构建模糊隶属度函数存在不足,以及易受数据集不平衡的影响,提出了一种基于高斯分布的FSVM,该方法既考虑了2类样本数量的不平衡问题,同时进一步考虑了样本不同方向上的分布特性.将样本的分布特性应用于模糊隶属度函数的设计,有效地提高了对正常样本和噪声、野值样本的区分能力.实验结果表明,在处理不平衡和有噪声干扰的数据集时,该方法较传统的FSVM具有更强的鲁棒性.

A FSVM for the imbalanced dataset based on the Gaussian distribution

Due to the defects in the fuzzy membership as a function of the distance of the sample points from the cluster center and its sensitivity to the imbalanced dataset in current fuzzy support vector machines( FSVM),a novel FSVM based on the Gaussian distribution is proposed. In the proposed method,the fuzzy membership is defined by not only the imbalance ratio between the numbers of two classes of datasets,but also those probability distributions. It can effectively improve the discrimination ability between normal samples with noise or outliers. Experimental results show that the proposed method is more robust than the traditional FSVM for the imbalanced datasets with noise.