Privacy Preserving Naïve Bayes Classification Using Trusted Third Party Computation over Distributed Progressive Databases

Abstract

Privacy-preservation in distributed progressive databases is an active area of research in recent years. In a typical scenario, multiple parties may wish to collaborate to extract interesting global information such as class labels without revealing their respective data to each other. This may be particularly useful in applications such as customer retention, medical research etc. In the proposed work, we aim to develop a global classification model based on the Naïve Bayes classification scheme. The Naïve Bayes classification has been used because of its simplicity, high efficiency. For privacy-preservation of the data, the concept of trusted third party with two offsets has been used. The data is first anonymized at local party end and then the aggregation and global classification is done at the trusted third party. The proposed algorithms address various types of fragmentation schemes such as horizontal, vertical and arbitrary distribution required format. The car-evaluation dataset is used to test the effectiveness of proposed algorithms.