摘要
数据通信业务收入发展状况受多种不确定因素影响,其预测过程是一个复杂的非线性问题,采用传统预测方法难以逼近复杂的非线性映射关系,其预测误差大.为此根据神经网络的非线性特性,提出一种基于RBF神经网络的数据通信业务发展预测模型,对2006-2011年数据通信业务收入发展进行预测,与BP网络训练结果比较,其收敛速度快,训练精度高,具有较强的鲁棒性和容错性,应用效果显著.
Data communication service income is influenced by many uncertain factors, its forcast process is a complex nonlinear problem, it is difficult to approach complex nonlinear mapping relation by traditional forcast method, and the forcast error is great. Therefore, according to nonlinear characteristic of neural network, a forecast model of data communication service income development based on RBF Neural Network is presented. It is applied in data communication service income forecast of 2006-2011, the result shows that it has faster convergence rate, higher training precision, stronger robustness, fault tolerance and more notable application effect than BP network.
出处
《邵阳学院学报(自然科学版)》
2007年第3期3-7,共5页
Journal of Shaoyang University:Natural Science Edition
基金
2006年国家自然科学基金项目(60377020)部分成果
关键词
RBF神经网络
数据通信业务收入
预测
RBF neural network
data communication service income
forecast