Research and application of network fault prediction algorithm
This study aims to investigate different network fault prediction techniques and use mathematical methods to construct models for transient analysis.We evaluate the performance of these techniques in predicting fault interval time using normalized root mean square error.The study employed multiple models and conducted continuous time Markov chain analysis to provide additional insights into network reliability.The results show that deep neural networks and autoencoders exhibit significant advantages in predictive performance,while linear models have poor performance,indicating that network fault data is more complex and not suitable for linear modeling.The time consumption of the network in suboptimal state is relatively high,and more proactive methods are needed to reduce the number of failures.A network fault prediction model based on continuous time Markov chains was utilized,and its effectiveness was demonstrated through experiments.
network faultprediction algorithmresearch and application
NETL
NSTL
万方数据
