Research Of Blockchain Security Transaction Based On On-chain Data

As an emerging technology,blockchain has the characteristics of decentralization,open and transparent transactions,and data tamper resistance,which can effectively improve data security and reduce data management costs.Today,blockchain technology has been applied to supply chain,financial technology,product traceability,public governance and other fields,and its development prospects are bright.Blockchain technology stores data through transactions between users.Users reach consensus in the blockchain system mainly rely on the consensus mechanism.However,there are criminals in the system who use methods such as blocking consensus to disrupt the normal transaction order of the blockchain.Due to the openness of data,some users have caused great security risks to transactions through fraudulent behaviors.The above transaction security issues have seriously affected the development and application of blockchain technology.Therefore,it is of great practical significance to study blockchain security transactions.This paper aims to ensure the security of blockchain transactions.Aiming at transaction security issues such as blocking consensus and fraudulent accounts on the chain,a research on blockchain security transactions based on on-chain data is carried out.The main research work is as follows:(1)Aiming at the security problem of malicious behavior of the master node in the practical Byzantine fault-tolerant consensus mechanism in the alliance chain,this paper proposes a consensus optimization scheme based on the reputation model based on the data structure on the chain.The scheme uses the P2 P reputation evaluation model to evaluate the actions of each node on the chain in the consensus process,and adds a node hierarchy system and a reputation incentive system to realize the real-time update of the reputation of each node on the chain and ensure that the nodes are more in the consensus.positive.Finally,through simulation experiments,the security and feasibility of the optimization scheme are analyzed in terms of fault tolerance,system stability and recoverability,consensus efficiency and system throughput.The results show that the scheme has better performance in security and feasibility.At the same time,the optimization scheme is compared with the existing optimization scheme,and the advantages and disadvantages of the scheme are compared and analyzed.(2)Aiming at the problem that the existence of fraudulent accounts on the blockchain causes normal user security transaction risks,this paper proposes an onchain fraud account detection scheme based on the data content on the chain,combined with the XGboost algorithm model and the SHAP model.The scheme firstly constructs the data set of the model by performing data preprocessing and feature engineering on the acquired real data on the Ethereum chain,and then constructs a fraudulent account detection model based on the XGboost algorithm and uses the real data set of Ethereum to verify the model.Compared with other machine learning methods,the results show that the detection model performs better,reaching 0.205,0.084 and 0.833 in the evaluation indicators such as RMSE,MAE and R2,respectively.Then,the grid parameter search technology is used to optimize the model.The performance of the model is further improved.Finally,the SHAP value model is introduced,and the key features of fraudulent accounts are discussed through feature analysis and feature importance comparison.According to the experimental results,it is shown that the consensus mechanism optimization scheme performs better than the original scheme,which can effectively reduce the malicious behavior of the master node in the practical Byzantine faulttolerant consensus mechanism,and improve the security and efficiency of users in transactions.At the same time,the fraudulent account detection scheme can be Effectively detect fraudulent accounts on Ethereum,and the key characteristics of fraudulent accounts obtained can provide reference when users conduct transactions.