Multi-Key Clustering Method for Cloud Environments' Privacy-Preserving

Data clustering is a crucial tool for data analysis, and it is reasonably efficient to outsource computational work to cloud servers in real-world applications where clients providing sample data frequently have CPU and storage capacity restrictions. Sadly, there is frequently a risk of customer data leakage when computing is outsourced. We provide a novel multi-secret K-Means clustering outsourcing strategy to protect privacy in cloud computing. The data owner separately encrypts sensitive data with the use of a distributed double-trap public key cryptosystem (DT-PKC) before uploading it to the cloud server. Then, using their respective secret keys, the cloud server and the service requester jointly compute K-Means clustering on the encrypted data. This system uses safe multiplication, secure squaring, and secure comparison protocols to support the underlying ciphertext operations. A secure Euclidean distance protocol for mutual cooperation is developed to safely calculate distances to update the set of prime numbers. The method effectively safeguards data privacy from all phases of the data source, calculation process, and output outcomes based on security analysis and time complexity analysis. The amount of computation and communication overhead is greatly reduced because the data owner only needs to communicate with the cloud once. Performance analysis demonstrates that our solution is both more beneficial than competing ones and functionally and safely similar to those already in use.