基于BP算法的多层前馈网络在质差建模中的应用  

Application of Multilayer Feed-Forward Network Based on BP Algorithm in Rxquality Modeling

作　　者：周智洪[1] 沙胜 

机构地区：[1]中国移动通信集团广东有限公司东莞分公司,广东东莞523129

出　　处：《移动通信》2015年第6期42-45,共4页Mobile Communications

摘　　要：BP神经网络算法可以以任意精度逼近任何非线性函数,且具有并行分布存储、高度鲁棒性与容错能力,适合解决受多重因素交叉影响的复杂的非线性问题。主要提出了一种基于BP神经网络的预测方法,并详细阐述神经算法的基本原理、建模方法以及其在质差预测中的应用。As BP neural network can approximate all nonlinear functions with arbitrary precision, has the abilities in parallel distributed storage, high robustness and fault tolerance, it appropriate for solving complex non-linear problem affected by multiple factors. In this paper, a prediction method based on BP neural network is presented. Then, its principle, modeling process and application in Rxquality prediction are expounded.

关 键 词：神经网络 BP算法 质差 

分 类 号：TN929.53[电子电信—通信与信息系统]