Research On Anonymous Traceable Method For Account Balance Transaction Model Blockchain

In recent years,the decentralization and anonymity of blockchain have given rise to an increasing number of illegal transactions based on virtual currencies.Therefore,how to achieve de-anonymization of blockchain addresses and transaction tracking,enabling effective regulation of blockchain transactions by the state,is a key issue for the further development of blockchain technology.However,through the analysis of existing research,the following problems are found: 1)Current research mainly focuses on the UTXO model blockchain represented by Bitcoin,and relatively less research on the anonymity and traceability of account balance model blockchain such as Ethereum;2)For large-scale data like Ethereum,traditional machine learning methods require complex feature engineering,and graph convolutional network methods need to take the entire graph structure as training input,consuming a large amount of resources and time with lower efficiency and accuracy;3)Existing transaction tracking research relies on expert experience and is only applicable to specific scenarios,making it difficult to effectively track historical transaction data and newly registered addresses.Therefore,this paper focuses on the research of de-anonymization and transaction tracking problems in account balance model blockchain,with the main work content as follows:(1)A Ethereum transaction address classification model based on AJK-Graph SAGE is proposed.The Ethereum transaction data is modeled as a graph structure,and the AJK-Graph SAGE algorithm is used to learn the embedding of the graph.The input of the entire model only requires the nodes and their sampled neighbor node sets.An attention mechanism is introduced,which assigns weights to each neighbor node based on strategies such as transaction frequency and transaction amount,making the contribution of each node to the output different.Secondly,different layers of feature vectors are aggregated through the jumping knowledge connection mechanism,making full use of the information of different layers,and improving the expressive power of the model.Finally,experiments were conducted to verify the accuracy and effectiveness of the model method.(2)A blockchain transaction tracking method based on ABW-Leader Rank algorithm for account models is proposed.In response to the problem of large-scale Ethereum transaction data and low algorithm running efficiency,a transaction subgraph construction method based on time segments is introduced.Then,the ABW-Leader Rank algorithm is used to iteratively calculate the influence of address nodes in the transaction subgraph,fully considering the influence of the number of transactions and transaction amounts on node influence,and fully considering the address nodes where the in-degree is much larger than the out-degree in the distribution strategy of the background node LR value after the algorithm converges.Finally,two types of graph search strategies are used to construct the path reachable set and the LR value set,and the proposed method is experimentally evaluated to prove its effectiveness.(3)Based on the key technology research and design proposed above,an account balance model blockchain anonymous traceable prototype system is designed and proposed on this basis.The design and implementation of various modules of the system are completed.Finally,the system is functionally tested to verify that the client and the server can provide services as expected.