Federated Learning Model for Auto Insurance Rate Setting Based on Tweedie Distribution

In the assessment of car insurance claims,the claim rate for car insurance presents a highly skewed probability distribution,which is typically modeled using Tweedie distribution.The traditional approach to obtaining the Tweedie regression model involves training on a centralized dataset,when the data is provided by multiple parties,training a privacy-preserving Tweedie regression model without exchanging raw data becomes a challenge.To address this issue,this study introduces a novel vertical federated learning-based Tweedie regression algorithm for multi-party auto insurance rate setting in data silos.The algorithm can keep sensitive data locally and uses privacy-preserving techniques to achieve intersection operations between the two parties holding the data.After determining which entities are shared,the participants train the model locally using the shared entity data to obtain the local generalized linear model intermediate parameters.The homomorphic encryption algorithms are introduced to interact with and update the model intermediate parameters to collaboratively complete the joint training of the car insurance rate-setting model.Performance tests on two publicly available datasets show that the proposed federated Tweedie regression algorithm can effectively generate Tweedie regression models that leverage the value of data fromboth partieswithout exchanging data.The assessment results of the scheme approach those of the Tweedie regressionmodel learned fromcentralized data,and outperformthe Tweedie regressionmodel learned independently by a single party.

Blockchain-Based Key Management Scheme Using Rational Secret Sharing

Traditional blockchain key management schemes store private keys in the same location,which can easily lead to security issues such as a single point of failure.Therefore,decentralized threshold key management schemes have become a research focus for blockchain private key protection.The security of private keys for blockchain user wallet is highly related to user identity authentication and digital asset security.The threshold blockchain private key management schemes based on verifiable secret sharing have made some progress,but these schemes do not consider participants’self-interested behavior,and require trusted nodes to keep private key fragments,resulting in a narrow application scope and low deployment efficiency,which cannot meet the needs of personal wallet private key escrow and recovery in public blockchains.We design a private key management scheme based on rational secret sharing that considers the self-interest of participants in secret sharing protocols,and constrains the behavior of rational participants through reasonable mechanism design,making it more suitable in distributed scenarios such as the public blockchain.The proposed scheme achieves the escrow and recovery of personal wallet private keys without the participation of trusted nodes,and simulate its implementation on smart contracts.Compared to other existing threshold wallet solutions and keymanagement schemes based on password-protected secret sharing(PPSS),the proposed scheme has a wide range of applications,verifiable private key recovery,low communication overhead,higher computational efficiency when users perform one-time multi-key escrow,no need for trusted nodes,and personal rational constraints and anti-collusion attack capabilities.