基于主成分分析降维和支持向量机回归的短期负荷预测方法研究

Short-term Load Forecasting Method Based on PCA and SVR

针对负荷预测一般使用多维度历史相关数据,随着计算机技术和采集技术的发展,形成了高维度数据,使得关键数据和因素被掩埋,造成“维数灾难”的状况,重点研究主成分分析的特征降维方法和支持向量机的智能预测算法,提出基于主成分分析和支持向量机回归的短期负荷预测方法,并进行算例计算,对比分析采用主成分分析降维对支持向量机短期负荷预测方法的影响。

Short-term power load forecasting is an important means of service power dispatching and control,as well as an important basis for power market transactions.Load forecasting generally uses multi-dimensional historical correlation data.With the rapid development of computer technology and acquisition technology,high-dimensional data is formed,which makes key data and factors buried,resulting in"dimensional disaster".Focusing on the feature dimension reduction method of principal component analysis and the intelligent prediction algorithm based on sup-port vector machine,a short-term load forecasting method based on principal component analysis and support vector is proposed,and the calculation is carried out.The influence of principal component analysis dimension reduction on the support vector machine short-term load forecasting method is analyzed and the applicable scope of the forecast-ing method is discussed.