Mobile agent-based frequent pattern mining for distributed databases

Abstract

In today’s world of globalization, business organizations produce information from many branch offices of their business while operating across the globe and hence lead to large chunk of distributed databases. There is an innate need to look at this distributed information that leverages the past, monitors the present, and predicts the future with accuracy. Mining large distributed databases using client–server model is time-consuming and sometimes impractical because it requires huge databases to be transferred over very long distances. Mobile agent technology is a promising alternative that addresses the issues of client–server computing model. In this paper, we have proposed an algorithm called MADFPM for frequent pattern mining of distributed databases that use mobile agents. We have shown that the performance of MADFPM is better compared to the conventional client–server approach.