基于GMDH-PSO-LSSVM中长期电力负荷预测

摘要

Aiming at the problem that the input node of the Least Squares Support Vector Machine model is difficult to be determined. A method based on Group Method of Data Handling-Improved Particle Swarm Optimization-Least Squares Support Vector Machine was proposed to solve the problem in medium and long term forecasting of power load. The specific method is as follows:firstly, the GMDH algorithm was used to obtain the input variables of the LSSVM modeling. Secondly, the adaptive mutation PSO algorithm was analyzed to optimize the parameters of the LSSVM model, and then the trained LSSVM model was utilized to predict the test samples. Furthermore, the actual load of a certain city from the year 2008 to 2013 was selected as the training samples of the model to establish the model, and the power load in 2014, 2015 were forecast by the trained GMDH-PSO-LSSVM model. The prediction results of the combined model show that the method achieves high prediction accuracy and the prediction accuracy is improved by 2.21%.%针对电力负荷预测粒子群优化最小二乘支持向量机(Least Squares Support Vector Ma-chine,LSSVM)模型输入节点难以确定的问题,提出了一种基于数据分组处理方法(Group Method of Data Handling,GMDH)来优化PSO-LSSVM(Particle Swarm Optimization-Least Squares Support Vector Machine)的中长期电力负荷预测预测方法.该方法是首先利用GMDH算法获得LSSVM建模中的输入变量;然后利用基于自适应变异的P SO算法对LSSVM建模中的参数进行优化,选用某地区2008~2013年的历史数据作为模型的训练样本建立模型;最后使用训练好的GMDH-P SO-LSSVM模型对2014、2015年的用电量进行外推预测.组合模型预测结果表明该方法达到了较高的预测精度,预测精度提高了2.21%.