Network Fault Root Cause Analysis Algorithm Based on Deep Learning

Abstract

With the rapid development of communication network, the number of network connections shows an explosive growth trend, and the number of network faults also increases. The existing intelligent control process is generally short of experience, which is still far away from the requirements of autonomous driving network. In this paper, knowledge graph, one of the key enabling technologies of autonomous driving network, is introduced into the field of fault root cause analysis, and the lack of experience is overcome by the priori nature of knowledge. This paper proposes a knowledge graph construction process for transport network fault root cause analysis. With its powerful knowledge management and the ability to fully mine the correlation between data, it has a great help to the network intelligent operation and maintenance. At the same time, some studies show that GNN algorithm can be used for entity and relationship reasoning of knowledge graph. This paper proposes a fault root cause analysis algorithm based on GGNN and knowledge graph, which uses GGNN to propagate aggregated information in the graph structure of fault graph to train fault location model. It will further get rid of the dependence on operation and maintenance personnel, reduce the operation and maintenance threshold and improve the efficiency of network management and control.