Privacy-Preserving Minority Oversampling Protocols with Fully Homomorphic Encryption

In recent years, blockchain and machine-learning techniques have received increasing attention both in theoretical practical aspects. However, the applications of these many challenges, one which is privacy-preserving issue. this paper, we focus on, specifically, issue imbalanced datasets, a commonly found problem real-world applications. Built based on fully homomorphic encryption technique, paper presents two new secure protocols, Privacy-Preserving Synthetic Minority Oversampling Protocol (PPSMOS) Borderline (Borderline-PPSMOS). Our analysis reveals that PPSMOS generally more efficient performance than Borderline-PPSMOS. Borderline-PPSMOS achieves better TP rate F-Value PPSMOS.