SYSTEM AND METHOD FOR DISCRIMINATING CORRECT OR NOT AND VERIFICATION INTEGRITY OF CYBER ATTACK PACKET DATA BASED ON MACHINE-LEARNING

摘要

The method of determining the mid-authenticity of the machine learning-based cyber attack packet data and verifying the integrity according to the present invention is (a) determining the mid-authenticity of the machine learning-based cyber attack packet data, and the transceiver of the integrity verification system communicates with a plurality of external systems. receiving packet data from a connected public institution system; (b) determining, by the main controller, whether the packet data received by the transceiver is existing data or new data; (c) if the data received by the transceiver in step (b) is new data, performing a new initial learning step by the main controller; And (d) if the data received by the transceiver in step (b) is existing data, performing a model re-learning step by the main controller; It has the effect of efficiently analyzing by applying an intelligent algorithm, and by storing the first packet in the block chain, it has the effect of verifying the integrity even if a hacker or an internal accomplice maliciously modifies the packet information. When the data integrity is verified by comparing the packets used and the packets stored in the block chain, model training is performed, thereby increasing the reliability of the model results.