A Computation of Frequent Itemset using Matrix Based Apriori Algorithm

Abstract

gunasekaran.cse@drmgrdu.ac.in
 The Apriori Algorithm is a traditional method for determining the frequent itemsets from a lot of data. Association rules can be generated based on frequently occurring item sets. The Apriori algorithm has two bottlenecks, namely that it generates a large number of candidate sets and that it repeatedly examines the database. It takes a long time to execute and takes up a lot of space. We provide a novel strategy called Matrix Based Apriori Algorithm to get beyond the limitations of Apriori Algorithm. We don't need to constantly scan the database because all operations are first applied to the matrix after which the database is converted back into its original form. In addition, we have reduced the potential itemsets by using several pruning techniques. The Matrix Based Apriori algorithm outperforms the standard Apriori algorithm in terms of time, with an average rate of time reduction of 71.5% with the first experiment and 86% with the second. In a similar vein, we contrasted our Matrix Based Apriori with an effective alternative known as improved apriori. We discovered that our method outperforms the upgraded apriori by 20%.