一种改进的FLS-SVM分类辨识模型及其应用

发布时间：2018-03-20 08:00

本文选题：混沌免疫算法　切入点：模糊最小二乘支持向量机　出处：《中南大学学报(自然科学版)》2017年08期 　论文类型：期刊论文

【摘要】：采用三角形函数隶属度法确定模糊最小二乘支持向量机(fuzzy least squares support vector machine,FLS-SVM)输入参数隶属度,采用自适应变尺度混沌免疫算法优化FLS-SVM的参数,从而构建改进模糊最小二乘支持向量机(improved fuzzy least squares support vector machines,IFLS-SVM)分类辨识模型,用Ripley数据集、MONK数据集和PIMA数据集进行仿真实验,并用于地下金属矿山采场信号分类辨识与中国国际贸易安全分类辨识。研究结果表明:与LS-SVM分类辨识模型和FLS-SVM分类辨识模型相比,IFLS-SVM分类辨识模型能有效提高带噪声点和异常点数据集的分类精度,且分类辨识精度相对误差较小。
[Abstract]:The membership degree of fuzzy least squares support vector machine input parameters is determined by triangular function membership method, and the parameters of FLS-SVM are optimized by adaptive variable scale chaotic immune algorithm. Thus, an improved fuzzy least squares support vector machine (fuzzy least squares support vector machine IFLS-SVM) classification identification model is constructed, and the simulation experiments are carried out with the Ripley data set and the PIMA data set. The research results show that compared with LS-SVM classification model and FLS-SVM classification identification model, IFLS-SVM classification identification model can effectively improve noise. Classification accuracy of acoustic and outlier data sets, And the relative error of classification and identification accuracy is small.
【作者单位】：中南大学资源与安全工程学院;湖南涉外经济学院商学院;
【基金】：国家自然科学基金资助项目(71573082) 湖南省自然科学基金资助项目(2017JJ2134) 湖南省高校创新平台开放基金资助项目(14K055)~~
【分类号】：TP18

【相似文献】