Research On Privacy Protection Of Supply Chain Financial Blockchain Oriented To Big Data Analysis

Supply chain finance involves a large number of users from all parties involved,including suppliers and purchasers and other small and medium-sized financing companies,core companies and financial institutions,and logistics companies.Supply chain financial blockchain faces many problems in the process of implementation and application.In the single-chain blockchain,supply chain financial business data cannot be isolated,which leads to node data redundancy,system failure for parallel processing and data privacy security threats.In addition,the established supply chain financial blockchain system has complex business,and its data is characterized by large scale and high dimension.In particular,banks and other financial institutions have a large amount of sensitive data and a high risk of disclosure of privacy information.Meanwhile,they have certain requirements on the availability of data for big data analysis.At present,some scholars have proposed some solutions for the privacy protection of blockchain,but most of the studies are about the privacy protection of identity privacy and transaction privacy information in the process of digital currency trading,and few studies have been conducted on the privacy protection of blockchain for big data analysis.Therefore,the first step is to extend the blockchain to find a solution that can solve the system efficiency deficiency in the single-chain structure.Secondly,it is necessary to study the privacy protection methods that can solve the privacy security problems and ensure certain data utility in the process of supply chain financial big data analysis.In view of the above mentioned problems,this paper mainly discusses the system efficiency of supply chain financial blockchain and the privacy protection in the process of big data analysis.Firstly,this paper studies the implementation of more differential privacy protection algorithms for machine learning and statistical applications in the industry,and proposes a differential privacy algorithm using feature extraction and Gaussian mechanism under high-dimensional data.Firstly,it is theoretically proved that the algorithm can satisfy the relaxed differential privacy,and the noise amount of the introduced data set is analyzed by noise variance.Then,the Frobenius norm and logistic regression classification prediction were calculated by setting experiments,and the data utility was compared with other related methods.Finally,the results show that the proposed method can reduce the introduced noise and improve the data utility by about 8% compared with other methods for data sets with high-dimensional data characteristics under the condition of satisfying the relaxed differential privacy protection.Then,a community clustering method is adopted to divide the nodes into different data-chains,and the data is isolated through multi-chain structure to solve the data privacy security threat and node data redundancy existing in the single-chain blockchain system.Through the performance analysis experiment,it is verified that the multi-chain blockchain has the capability of parallel processing requests,which can reduce the response time of the system and improve the efficiency of processing requests.Finally,in view of the query limitation caused by excessive privacy budget loss,this paper proposes a noise multiplexing differential privacy algorithm,and verifies the privacy budget loss of the algorithm and the utility of data results through experiments.The results show that the proposed algorithm can record the usage of the privacy budget of the blockchain system through the historical record ledger,and by reusing the old noise,it can effectively reduce the loss of the privacy budget and prevent the query limitation caused by the waste of the privacy budget.