SYSTEM AND METHOD FOR NETWORK FAULT DETECTION

The disclosure relates to techniques for detecting network faults. Such techniques may include obtaining an input data set from a plurality of nodes of a network asset and predicting a faulty node of the plurality of nodes by inputting the input data set to a class of classifiers. The class of classifiers may be trained according to normal operation data obtained during normal operation of the network asset. Further, the network fault detection technique may include calculating a confidence level for network fault detection of the input data set using the class of classifiers, and adjusting a decision threshold based on the confidence level to classify the input data set as normal or including a network fault. The predicted faulty node and the adjusted decision threshold may be used to detect a network fault in the plurality of nodes. 本公开涉及用于检测网络故障的技术。此类技术可以包括从网络资产的多个节点获得输入数据集，并且通过将所述输入数据集输入到一类分类器来预测所述多个节点中的故障节点。所述一类分类器可以是根据在所述网络资产的正常操作期间获得的正常操作数据进行训练的。此外，所述网络故障检测技术可以包括使用所述一类分类器来计算所述输入数据集的网络故障检测的置信水平，并且基于所述置信