The K-Anonymity Approach in Preserving the Privacy of E-Services that Implement Data Mining

In this paper, we first described the concept of k-anonymity and different approaches of its implementation, by formalizing the main theoretical notions. Afterwards, we have analyzed, based on a practical example, how the k-anonymity approach applies to the data-mining process in order to protect the identity and privacy of clients to whom the data refers. We have presented the most important techniques and algorithms used in order to enforce kanonymity. We have studied possible approaches to ensure that k-anonymity preserves the privacy of the data mining process in order to assure an efficient, safe and effective data mining delivery as an e-service. We have depicted several software implementations that employ k-anonymity. We have developed, using the Compute Unified Device Architecture, an algorithm that analyzes k-anonymity and is suitable for extracting tuples with the same quasiidentifying values from a large database structured as a private table.