Privacy Preserving Three-Layer Naïve Bayes Classifier for Vertically Partitioned Databases

Abstract

Data mining is the extraction of the hidden information from large databases. It is a powerful technology to explore important information in the data warehouse. Privacy preservation is a significant problem in the field of data mining. It is more challenging when data is distributed among different parties. In this paper, we address the problem of privacy preserving three-layer Na\u00efve Bayes classification over vertically partitioned data. Our approach is based on Secure Multiparty Computation (SMC). We use secure multiplication protocol to classify the new tuples. In our protocol, secure multiplication protocol allows to meet privacy constraints and achieve acceptable performance and our classification system is very efficient in term of computation and communication cost.