A Survey on Privacy Preserving Data Mining Techniques

Objectives: Recently Privacy preserving data mining (PPDM) is known to be the most important aspect among researchers. As Privacy preserving data mining permits, sharing and exchanging of privacy susceptible data for analysis, it has grown more and more popular. Since one of the important aspects of data mining is safeguarding privacy, this paper aims to analyze different technique adopted for preserving privacy while maintaining the real characteristic of data under consideration. Methods/Statistical Analysis: In this paper, the authors evaluate the usefulness of PPDM techniques based on its performance, data usage, and uncertainty level and so on. The findings of authors and limitations in each technique are consolidated. Findings: Each technique has its unique way of usefulness apart from its limitations. Anonymization approach makes the data owners anonymous but vulnerable to attacks like linking attacks. Perturbation approach protects each and every attribute independently but unable to regenerate the original values from the perturbed data. Randomization technique provides good security for individual’s private data but the utility of the data. The degradation of the utility of the data is due to the noise added. The cryptographic technique provides good security for the data while providing high utility. But it falls short in efficiency when compared with other methods. Anyhow, there is no single privacy protecting algorithm capable of outperforming every other algorithm in all possible yardsticks. On the contrary, one algorithm may do well when compared to another, on a particular criterion. Novelty/Improvement: The paper presents various techniques which are used to perform PPDM technique and also tabulates their advantages and disadvantages.