Homomorphic Encryption Based Privacy Protection Scheme Design for Tourism Big Data

With the rapid development of tourism big data, data privacy and security issues are becoming more and more prominent. Traditional encryption technology is risky due to the need to decrypt the ciphertext, while full homomorphic encryption is difficult to be applied in practice due to high computation and storage costs. In this paper, we propose a privacy protection scheme based on homomorphic encryption for tourism big data, which combines federated learning and blockchain dynamic authorisation mechanism with dynamic encryption layer selection algorithm to optimise privacy protection and computational efficiency. The scheme adopts BFV homomorphic encryption algorithm, optimises the key generation process (the key size is compressed to 45MB), and integrates Microsoft SEAL library to build an efficient homomorphic computing engine. The solution supports high concurrent processing of ten million data in hybrid cloud environment, privacy protection is increased, and the loss of analytical accuracy of encrypted data in SVM and XGBoost model is reduced by 4.1%. GDPR compliance requirements are met by introducing spatio-temporal indexed encryption with lightweight key proxy services. The research provides a secure and efficient solution for multimodal data analysis in tourism scenarios, which will be further combined with hardware acceleration and standardised integration in the future to promote privacy-preserving practices in smart tourism.