基于人工智能和大数据的网络故障诊断可视化平台研究

Research on Network Fault Diagnosis Visualization Platform Based on Artificial Intelligence and Big Data

异构网络内部结构的复杂性增加了网络故障诊断的难度,为得到更加准确高效的网络故障诊断模型,研究构建以图卷积神经网络为基础的网络故障诊断可视化平台,通过生成对抗网络保证网络数据集的平衡性,利用朴素贝叶斯生成拓扑关联图。实验结果表明,该模型的准确率稳定在92%左右,且其对故障的整体判别正确率达到了92.5%,比改进前提升了2.4%。通过实验证明研究提出的改进图卷积神经网络算法能提高网络故障检测的精度,为网络通讯技术在各领域的普及提供了良好的借鉴。

The complexity of the internal structure of heterogeneous networks increases the difficulty of network fault diagnosis,in order to obtain a more accurate and efficient network fault diagnosis model,the research constructs a network fault diagnosis visualisation platform based on graph convolutional neural network,ensures the balance of the network dataset by generating an adversarial network,and generates the topological correlation graph using plain Bayes.The experimental results show that the accuracy of the model is stable at around 92%,and its overall correct rate of fault discrimination reaches 92.5%,which is 2.4%higher than that before the improvement.It is proved experimentally that the improved graph convolutional neural network algorithm proposed in the study can improve the accuracy of network fault detection,which provides a good reference for the popularity of network communication technology in various fields.