Data mining is a powerful tool to extract information and find patterns from huge dataset; meantime data mining also pose a threat to the privacy and security issues. Privacy preserving data mining is the key for this; it applies data mining technique while considers various privacy and security constrains. Privacy preserving data mining has been intensively investigated in recent years, and mainly can be classified into two categories, perturbation and randomization approaches, and secure multi-party computation (SMC). Our research more focuses on the first one, perturbation and randomization technique. We have conducted experiments both on Bayesian based and Principal Components Analysis (PCA) approaches, and furthermore have proposed an adaptable model which will allow individual to choose his or her own privacy level hence perturbation level.
