摘要
研究网络故障诊断问题,保证网络可靠性运行效率,针对网络故障是一个非线性、小样本数据,但是传统网络故障诊断方法是基于线性、大样本数据,导致网络故障诊断准确率较低。为了提高网络故障诊断准确率,将专门解决小样本、非线性问题的最小二乘支持向量机(LSSVM)应用到网络故障诊断中,将引起故障的因素作为LSSVM的输入,网络故障类型作为LSSVM输出,通过LSSVM的学习,建立网络故障诊断模型,最后采用建立的LSSVM模型对网络故障样本进行诊断。仿真结果表明,LSSVM网络故障诊断准确率明显高于其它网络故障诊断方法,并证明是一种网络故障诊断有效手段。
Research computer network fault problems.The computer network fault is of typical nonlinear and small sample data,and the traditional mathematical model and artificial intelligence method cannot obtain good resultd.In order to improve the accuracy of computer network fault diagnosis,we applied the least squares support vector machines（LSSVM） to computer network fault diagnosis field.LSSVM is an effective model for solve small sample amd nonlinear machine learning algorithm.We used the optimized LSSVM model for training and testingg.Simulation experiments showed that the diagnostic accuracy of LSSVM is higher than the BP neural network,and it is a computer network fault diagnosis model with high accuracy.
出处
《计算机仿真》
CSCD
北大核心
2011年第10期103-106,共4页
Computer Simulation
关键词
网络故障诊断
支持向量机
神经网络
Network fault diagnosis
Support vector machines（SVM）
Neural networks（NN）
