An alliance chain-based incentive mechanism for PSG data sharing

【Author】 Zhang, Wenbo; Huo, Xiaotong; Bao, Zhenshan

【Source】PEER-TO-PEER NETWORKING AND APPLICATIONS

【影响因子】3.488

【Abstract】Deep learning-based sleep staging can identify the signal characteristics of each stage of polysomnography (PSG) and classify them automatically, but the sleep staging learning model for pursuing higher classification accuracy cannot be separated from the aggregation of a large number of data sets. To meet the needs of decentralized medical institution deployment and data security, this paper proposes an alliance chain-based PSG data sharing model. Considering the need for PSG data to carry tags when sharing and the access of wearable IoT devices when collecting, we design an alliance chain-based intra- and inter-medical institution chain to ensure data security and privacy, and a labeled off-chain storage scheme based on the planetary file system (IPFS), which guarantees the blockchain performance and meets the need for labeling. At the same time, PSG is a kind of precious medical data, and it is difficult for each medical institution to share the dataset spontaneously without compensation. Therefore, this paper designs an incentive mechanism based on data value and access credit, and obtains the incentive evaluation value by evaluating the dataset quality and institutional access behavior through an open and transparent mechanism, which not only guarantees the activity of the sharing system but also restrains the access behavior of each institution. Finally, this paper uses Hyperledger Fabric to implement a resource access control model based on permission hierarchy through smart contracts, and narrows the scope of policy retrieval by introducing incentive evaluation values, and the model is tested under six different policy scales with different degrees of significant improvement in access efficiency.

【Keywords】Alliance chain; PSG data sharing; Incentive mechanism; Access contorl

【发表时间】2023 2023 OCT 21

【收录时间】2023-11-09

【文献类型】实验仿真

【主题类别】

区块链应用-实体经济-医疗领域