摘要
在电力行业信息化发展背景下,收集与存储大量电力数据,可为电力企业营销决策制定提供依据。该文提出采用人工神经网络构建电力营销系统BP神经网络模型,通过智能决策树分类算法预处理模型数据,得到最优化的模型数据,并改进神经网络隐含层节点数目算法,结合应用分时段预测方法及共轭梯度算法分别进行网络训练及网络结构优化,为网络收敛速度加快提供保障,得出相对准确的电力营销年度电量预测结论,说明电力营销系统中人工神经网络具有较高的应用价值。
In the context of the development of electric power industry informatization,a large number of electric power data are collected and stored,which can provide a basis for marketing decision-making of electric power enterprises.This paper proposes to use artificial neural network to construct BP neural network model of electric power marketing system,preprocess model data by intelligent decision tree classification algorithm,get the optimal model data,and improve the algorithm of hidden layer node number of neural network.Network training and structure optimization are carried out by using time-divided prediction method and conjugate gradient algorithm,which provides a guarantee for accelerating the convergence speed of the network.A relatively accurate conclusion of annual electricity forecast of electric power marketing is obtained,which shows that artificial neural network has high application value in electric power marketing system.
出处
《科技创新与应用》
2024年第13期167-170,共4页
Technology Innovation and Application
关键词
人工神经网络
电力营销
误差反向传播模型
BP神经网络模型
决策树分类算法
artificial neural network
power marketing
error back propagation model
BP neural network model
decision tree classifacation algorithm