Application Research of Graph Embedding Method Based on Feature Fusion in Ethereum Wallet Abnormal Data Recognition
In recent years,due to the strong privacy of Ethereum wallet data,it is difficult for attackers to detect such data anomalies in cryptocurrency transactions.This article studies the Ethereum illegal transaction recognition application based on feature fusion graph embedding method.This method includes two feature extraction strategies:abnormal data feature extraction and transaction feature extraction.Specifically,using BP neural networks to extract abnormal data features from Ethereum wallets,and using a random walk strategy to extract transaction features.Then,the extracted abnormal data features and transaction features are fused to obtain an Ethereum wallet abnormal data representation.The experimental results show that this method outperforms other algorithms in various indicators and can effectively detect abnormal data in Ethereum wallets.
ethereum walletabnormal transactionsgraph construction
NETL
NSTL
万方数据
