Proposed Parallel Association Rules Algorithm

Abstract

Data mining is an advanced technique for extracting knowledge from a large amount of data for classification, prediction, estimation, clustering or association rules or any activities, which need decision. Mining for associations rules between items in large transactional distributed databases is a central problem in the field of knowledge discovery. When distributed databases are merged at single machine to mining knowledge it will require a large capacity of storage, long execution time in addition to transferring a huge volume of data over network might take extremely long time and also require an unbearable financial cost. In this paper an algorithm is presented toward saving communication cost over the network, central storage cost requirements, and accelerating required execution time. In this paper a new algorithm is proposed, called Proposed Parallel Association Rules Algorithm (PPARA) which aims to extract association rules from one record only for each site from distributed association rules in parallel instead of extracting association rules from huge quantity of distributed data at several sites in parallel, and that is through collecting the one record of local association rules from each site and storing it, these Local Association Rules turn in to produce global association rules over distributed systems in parallel.