Intention identification and prevention method of abnormal transaction on a blockchain, storage medium and electronic device thereof

Abstract

The application discloses a method for identifying and preventing abnormal transactions on a blockchain, a storage medium and an electronic device. High-frequency transactions are common on the blockchain, and there are a large number of coincidental transactions that look abnormal but are not attack intentions on the chain. Existing methods mostly match shapes and do not distinguish intentions at the entity level, resulting in more false positives and lower precision after cross-block window expansion. The application obtains shape candidates composed of a pre-run set, a victim set and a post-run set, and satisfying the sorting window, amount matching and attacker association constraints. Four types of signals, including transaction package, economy, slip point consumption and pre-run position interval, are calculated after aggregation according to the attacker entity. When the transaction package signal is satisfied or the last three gates are passed at the same time, it is determined that the entity has an attack intention, thereby distinguishing between real attacks and coincidental transactions. Further, disposal such as rejection and sorting exclusion is performed on the malicious address at the gateway or sorting node to prevent subsequent abnormal transactions. The application can be run in real time block by block and significantly improve the precision.