因子分析和神经网络的光纤通信网络风险分析  

Risk Analysis of Optical Fiber Communication Network by Factor Analysis and Neural Network

作　　者：程凯[1] 董雪[1] 

机构地区：[1]河南教育学院,郑州450046

出　　处：《激光杂志》2016年第2期118-121,共4页Laser Journal

基　　金：河南省社会发展领域科技攻关项目(112102310377)

摘　　要：风险分析是光纤通信网络的关建技术,而网络风险的影响因子多、因子之间存在相互作用,为了提高光纤通信网络风险分析的准确性,发现存在潜在的网络风险,提出一种因子分析和神经网络的光纤通信网络风险预测模型(FA-RBF)。首先尽可能多的收集光纤通信网络影响因子,并采用因子分析法对影响因子进行处理,选择主要影响因子,然后根据主要因子确定RBF神经网络结构,采用遗传算法对RBF神经网络参数进行选择,最后的应用实例结果表明,FA-RBF可以快速度、准确确定光纤通信网络的影响因子,减少了RBF神经网络向量的维数,提高了光纤通信网络风险分析的精度,加快了风险分析的速度。Risk analysis is the key technology of optical fiber communication network,and there are many factors that affect the network risk,and the interaction among the factors,In order to improve the accuracy of risk analysis and find the potential network risk,a risk forecast model for optical fiber communication network and factor analysis is proposed in this paper.Firstly,the influencing factors of optical communication network are collected as more as possible,and factor analysis is used to determine the main influencing factors,and secondly RBF neural network structure is determined by the main influence factor,and the parameters of RBF neural network are selected by using genetic algorithm,lastly,the results show that the proposed model can accurately determine the influence factors of optical fiber communication network,reduce the dimension of RBF neural network,and improve the accuracy of risk analysis,and speed up the risk analysis.

关 键 词：光纤技术 通信介质 因子分析 神经网络 网络风险 

分 类 号：TN27[电子电信—物理电子学]