杨胡萍,左士伟,涂雨曦,王承飞.基于混沌理论和Legendre正交基神经网络的短期负荷预测[J].电测与仪表,2015,52(13):. YANG Hu-ping,ZUO Shi-wei,tuyuxi,WANG Cheng-fei.Short-term Load Forecasting Based on Chaos Theory and Legendre Orthogonal Basis Neural Network[J].Electrical Measurement & Instrumentation,2015,52(13):.
基于混沌理论和Legendre正交基神经网络的短期负荷预测
Short-term Load Forecasting Based on Chaos Theory and Legendre Orthogonal Basis Neural Network
Considering the short-term load’s characteristics of chaos and the neural network’s nonlinear mapping ability, this paper puts forward a kind of prediction method which is based on chaos theory and Legendre neural network. The method selects best training samples based on Euclidean distance after using chaos theory to reconstruct space for short-term load data, and then train the network using three layer neural networks whose excitation function of hidden layer is Legendre orthogonal polynomials,at last uses the trained network to forecast. In order to determine the network’s best topological structure when the network is training, this paper introduces derivative algorithm to determine the best number of hidden layer neurons. Example analysis shows that the method is feasible, and can get a high forecasting precision and good prediction effect.