Information Disclosure, Security, and Data Quality

Data sharing is important to interested parties for mining trends and patterns in designing data-driven decision-making systems. However, sharing raw data creates severe problems like identity theft or personal information leakage such as disclosure of illness of a specific person. This study analyzed results from an Adaptive Differential Privacy (ADiffP) algorithm that satisfies e-differential privacy for publishing sanitized data. The algorithm was tested with two different data sets to measure its robustness by comparing other existing works. The results obtained from the proposed algorithm show that the sanitized data preserves the same pattern as the raw data. Additionally, classification accuracies for sanitized data are also promising.