Blockchain log monitoring method based on deep learning

The invention discloses a blockchain log monitoring method based on deep learning. Under the assumption that the blockchain does not crash due to too many malicious nodes, a deep learning anomaly monitoring model is deployed on limited blockchain nodes to monitor logs according to a consensus mechanism and a fault-tolerant algorithm of the blockchain, and an anomaly alarm is given out if the logshave abnormal performance. Wherein in the aspect of model training of deep learning, on the basis of a model obtained through pre-training, the latest log evaluated as normal operation is added into the training of the model at set intervals so as to keep timeliness. The invention provides a log monitoring method capable of realizing automatic maintenance, which is used for monitoring abnormity ofan operation level of a block chain system in real time from a log level and giving an alarm. 本发明公开了一种基于深度学习的区块链日志监测方法，假设区块链没有因为恶意节点过多而崩溃的情况下，根据区块链的共识机制以及容错算法，在有限的区块链节点上部署深度学习异常监测模型来监测日志，如果日志存在异常表现则抛出异常报警。其中在深度学习的模型训练方面，在预训练得