摘要
利用小波变换的时频局部化特性和神经网络的自学习能力,构建基于小波变换的神经网络,从而使该神经网络具有较强的逼近能力和容错能力。将小波神经网络应用于入侵特征分类技术当中,具有了较强的自适应、自学习能力,并通过对动量项因子和学习因子的调整对模型进行改进,解决传统的前馈神经网络易陷入局部极小点的问题,加快神经网络的收敛速度。
According to the wavelet transform time-frequency localization characteristics and self-learning ability of neural networks,constructs wavelet neural network.So that the wavelet neural network has the strong approimation ability and fault tolerance.Applying the wavelet neural network to the intrusion features classification,it has a strong adaptive,self-learning ability.Through the momentum term adjustment factor and learning factor improves the model.It solves the problem of the traditional BP neural networks easy to fall into local minimum and speeds up the convergence rate of neural network.
出处
《现代计算机》
2011年第6期13-16,共4页
Modern Computer
关键词
小波神经网络
入侵特征分类
动量项因子
学习因子
Wavelet Neural Network
Intrusion Features Classification
Momentum Factor
Learning Factor
