机构地区: 广东省肇庆供电局肇庆526060
出 处: 《现代计算机》 2009年第8期17-21,共5页
摘 要: 构建基于RBF神经网络的电力客户信用评估模型。为了改善样本的质量和降低训练神经网络时对样本的处理成本,对原始样本进行主成分分析,利用提取的少数主成分作为新的综合指标再进行评估模型的构建,同时通过改进MatLab神经网络工具箱中的RBF神经网络设计函数,构建高分类能力的网络模型。实验证明该信用评估模型的有效性。 Builds a risk evaluation model for electricity customers based on the RBF neural networks. In order to improve the quality of samples and reduce the cost of processing samples while training the neural network, uses principal component analysis to draw a small number of principal components as new composite indicators before constructing the model. Also constructs a high-capacity neural network classification by improving the RBF neural network design function in the MatLab neural toolbox. Experimental results Drove the validity of the model.
领 域: [自动化与计算机技术] [自动化与计算机技术] [经济管理]