EDNBC: A New Efficient Distributed Naive Bayes Classifier for Vertically Distributed Data

A common constraint in distributed data is that the database cannot be moved to other network sites due to computational costs, data size, or privacy considerations. All of the existing distributed Naive Bayes algorithms for classifying data are designed for horizontally distributed or special cases of vertically distributed data where different sites contain different attributes for a common set of entities. In this paper, we propose a new distributed version of the Naive Bayes Classifier (EDNBC) using a Directed Acyclic Graph (DAG) in d-dimensional space across vertically distributed databases. The main goal of the proposed version is to minimize the communication cost among the database nodes by gathering statistical summaries at each site and then aggregating these summaries to get the final results.