基于主成分与粒子群算法的LS-SVM短期负荷预测

LS-SVM Short-term Load Forecasting Based on Principal Component Analysis and Improved Particle Swarm Optimization

短期电力负荷预测对电力系统安全经济运行和国民经济发展具有重要意义。最小二乘支持向量机(Least square support vector machines,LS-SVM)在解决小样本、非线性问题中表现出许多特有的优势,该方法已成功应用在负荷预测领域。本文提出了一种基于主成分分析的支持向量机预测模型,运用主成分分析对历史数据进行主成分提取,消除输入的训练数据本身存在着大量的噪声和冗余,从处理后的数据提取LSSVM的训练样本,并利用改进的粒子群优化算法(Particle Swarm Optimization,PSO)以LSSVM中的参数作为粒子进行优化,进而提高训练速度和预测精度。最后,将该模型运用到短期电力负荷预测中,与经典的SVM和BP神经网络相比具有更好的泛化性能和预测精度。

Short-term load forecasting is of great significance for power system economic operation and development of national economy. Least squares support vector machines （LSSVM） has been successfully applied in load forecasting, which has many unique advantages in the performance of solving the small sample, nonlinear problems. This paper presents a principal component analysis based on support vector machine model, using the principal component analysis to extract the principal components of historical data and eliminate a lot of noise and redundancy, then data extraction from the processed LSSVM training samples, and using improved particle swarm optimization which regards parameters in LSSVM as particles to improve the training speed and prediction accuracy. Finally, the model is applied to short term load forecasting, and has better generalization performance and prediction accuracy compared to SVM and BP neural network.