Efficient Tree Based Distributed Data Mining Algorithms for mining Frequent Patterns

Abstract

Advancements in the field of wired and wireless network environments have paved route to the advent of many dynamic distributed computing environments. These environments have diverged computing resources and multiple heterogeneous sources of data. Most mining algorithms are designed to mine rules from monolithic non-distributed databases. Even algorithms exclusively designed to operate on distributed databases normally download the relevant data to a centralized location and then perform the data mining operations. This centralized approach does not work well in many of the distributed, ubiquitous, privacy sensitive data mining applications, which opened a new area of research Distributed Data Mining (DDM) under the data mining domain. Out of various methods employed to mine frequent Itemsets, tree based methodology proves some efficiency in distributed environment. So in this paper we study a set of tree based algorithms [DTFIM, PP, LFP and PP] to mine frequent pattern in distributed environment.