基于PSOGA的LS-SVM模型在时间序列预测中的应用

Application of LS-SVM Based PSOGA Model to Time Series Prediction

摘要: 使用PSO与GA结合的混合算法PSOGA对最小二乘支持向量机(LS-SVM)模型的参数进行了优化,搜索到更优的参数,提高了模型的时间序列预测精度.在Mackey-Glass、Lorenz时间序列上的实验结果表明:本文模型预测精度较高.

Abstract: The LS-SVM model parameters were optimized by using PSOGA hybrid algorithm which combined PSO with GA，the better parameters were searched and the time series prediction accuracy of the model were improved.The experimental results on Mackey-Glass、Lorenz time series further verified the effectiveness of our model