粒子群优化LSSVM与RBF的混合回归预测模型  

Hybrid Regression Prediction Model Based on Particle Swarm Optimization LSSVM and RBF

作　　者：王博[1] 彭硕[1] WANG Bo;PENG Shuo(School of Electronic and Information Engineering,Jinggangshan University,Ji’an 343009,China)

机构地区：[1]井冈山大学电子与信息工程学院,江西吉安343009

出　　处：《软件导刊》2021年第12期49-53,共5页Software Guide

基　　金：国家自然科学基金项目(61862035);江西省教育厅科技计划项目(GJJ190561)。

摘　　要：为提高支持向量机预测模型的精确度,将最小二乘向量机与RBF神经网络引入预测模型,使用粒子群算法对RBF神经网络进行优化,使RBF神经网络能够提取向量机训练误差中的有效值,并再次进行数据集训练,使RBF神经网络模型能进一步预测支持向量机残差项。将粒子群算法优化的LSSVM-RBF混合回归预测模型用于大流域生态水环境预测,并将预测结果与其他4种模型进行对比。该模型的RMSE、MAE、MAPE、MSE分别达到了1.1077、0.9135、0.2500、1.2270,拟合程度进一步提高,表明基于粒子群算法优化的LSSVM-RBF⁃NN混合预测回归模型能够有效提高预测精度。In order to improve the accuracy of SVM prediction model,introduces least square vector machine and RBF neural network into the prediction model,and optimizes RFB neural network by particle swarm optimization,so that RBF neural network can extract the effective value of training error of vector machine and train the data set again,The RBF neural network model can further predict the residual term of SVM.LSSVM RBF mixed regression model optimized by particle swarm optimization was used to forecast the eco⁃logical water environment of the large watershed.Compared with the other 4 models,RMSE,MAE,MAPE and MSE reached 1.1077,0.9135,0.2500 and 1.2270 respectively,and the degree of fitting was further improved.the LSSVM RBFNN hybrid prediction regres⁃sion model based on particle swarm optimization can improve the prediction accuracy effectively.

关 键 词：粒子群算法 最小二乘支持向量机 RBF神经网络 回归预测 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]