Blockchain and deep learning technology for comprehensive improvement of transaction information quality

Due to big data technology's rapid development, the demand for data from related industries is growing, but most of the data is currently in a decentralized island state, and organizations that can release data value may need more data. Therefore, improving the information quality of transactions is very important for enterprises. Based on blockchain technology, this paper comprehensively enhances the quality of transaction information from three aspects: information security, information hiding, and data asset trading. The first part of this study develops an enhanced logistics data access control technique using attribute-based hierarchical ciphertext encryption. To secure the security of users' decryption keys, this technique substitutes the conventional key distribution centre with the authority management module in the super ledger. Secondly, to improve the concealment of blockchain-based information-hiding schemes, this paper uses an Recurrent Neural Network based self-encoder to generate secret transactions. It uses a Convolutional Neural Network discriminator to improve the concealment of secret transactions. Finally, we propose a data transaction scheme based on blockchain and trusted computing. Store data index information and transaction information in the blockchain and use the blockchain as a trusted storage environment for data transaction index through smart contracts. The experiment shows that our three comprehensive strategies achieve fine-grained access to data to ensure the privacy of users' data. It can embed secret information in Ethereum transactions, so it is robust and can effectively improve the comprehensive quality of transaction information.