针对支持向量机中由于噪声和野值带来的过拟合问题,提出了基于模糊最小二乘支持向量机(fuzzy LS-SVM)的抑制非直达波的移动定位方法。利用一种基于支持向量数据域描述的模糊隶属度函数模型,根据样本到特征空间最小包含超球球心的距离来确定模糊隶属度。仿真结果表明了该方法的稳健性,提高了LS-SVM的抗噪声能力。
In order to overcome the overfitting problem caused by noises and outliers in support vector machine, a method for non-line-of-sight (NLOS) mitigation based on fuzzy least square support vector machines (LS-SVM) is proposed. Using the fuzzy membership model based on support vector data description (SVDD), the membership values to each input sample is computed according to its distance to the center of the hypersphere with minimal volume containing all objects. Simulation results show that the proposed method is robust in NLOS environments and actually increases the accuracy of LS-SVM.