基于神经网络的话务量预测  

Mobile telephone traffic prediction model based on neural network

作　　者：邓波[1] 李健[1] 孙涛[1] 张金生[1] 王惠东[1] 

机构地区：[1]四川大学电子信息学院,四川成都610064

出　　处：《成都信息工程学院学报》2008年第5期518-521,共4页Journal of Chengdu University of Information Technology

摘　　要：话务量具有高度的非线性和时变特性,由于神经网络具有较强的非线性映射等特性,将其运用于非线性的话务量短期预测是非常合适的。以青白江2005年10月的话务量作为预测对象,提出基于BP神经网络和基于Elman神经网络的话务量预测模型,仿真实验表明两种模型对于话务量的短期预测均是可行有效的。经过比较,Elman神经网络训练速度比BP神经网络快很多,更适用于实际应用。The mobile tdephone traffic has typical time-varying and nonlinear features. With the nonlinear mapping feature the neural network is fit for the telephone traffic prediction. The Qingbaijiang tdephone traffic is used as the prediction sample. The BP neural network and Elman neural network am taken as the prediction models. The simulation results show that the two models are efficient to predict the mobile telephone traffic. But the Elman neural network has a comparative faster calculation rate than the BP neural network and is morn suitable for the practical application.

关 键 词：BP神经网络 ELMAN神经网络 话务量预测 预测模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]