Matlab Implementation of Parallel Apriori Algorithm for Reduced Space Time Complexity

Abstract

Data mining has gained wide spread usage & acceptance in the last decade. There are numerous tasks & fields in which data mining provides knowledge discovery solution one of the primary advents of data mining is determination of frequent of patterns or frequent patterns & their count. Another of data mining techniques are available for frequent pattern search but apriori algorithm is highly preferred & proved choice & caters to small as well as large data sets & database. This work is aimed at optimizing the time & space consumed by apriori algorithm without sacrificing the basic apriori construct by employing parallel computing services of matlab. Unlike other work carried to enhance apriori algorithm by modifying its iterations or traversal, our proposed system keeps the basic apriori proposal intact & enhance the apriori algorithm by using parallel computing to utilize all cores of the CPU or the network. The proposed work also support processing nodes spread over internet, thus various remote processing serves can handle users apriori search, thus improving user experience user modifier while monitoring apriori integrity.