This paper proposes a fuzzy-based Lagrangian twin parametric-margin support vector machine (FLTPMSVM) to reduce the effect of the outliers presented in biomedical data. The proposed FLTPMSVM assigns the weights to each data sample on the basis of fuzzy membership values to reduce the effect of outliers. This paper also adopts the square of the 2-norm of slack variables to make the objective function more convex. The proposed FLTPMSVM solves simple linearly convergent iterative schemes instead of solving a pair of quadratic programming problems. No external toolbox is required for the proposed FLTPMSVM as compared to the other methods. To establish the applicability of the proposed FLTPMSVM in the area of biomedical data classification, numerical experiments are performed on several biomedical datasets. The proposed FLTPMSVM gives an improved generalization performance and reduced training cost as compared to support vector machine (SVM), twin support vector machine (TWSVM), fuzzy twin support vector machine (FTSVM), twin parametric-margin support vector machine (TPMSVM) and new fuzzy twin support vector machine (NFTSVM).

本文提出了一种基于模糊的拉格朗日双参数裕度支持向量机（FLTPMSVM），以减少生物医学数据中异常值的影响。所提出的FLTPMSVM基于模糊隶属度值将权重分配给每个数据样本，以减少离群值的影响。本文还采用了松弛变量2-范数的平方，以使目标函数更凸。提出的FLTPMSVM解决了简单的线性收敛迭代方案，而不是解决一对二次规划问题。与其他方法相比，建议的FLTPMSVM不需要外部工具箱。为了建立所提出的FLTPMSVM在生物医学数据分类领域的适用性，对几个生物医学数据集进行了数值实验。