Association Rules Mining Exact, Approximate and Parallel Methods: A Survey

Association rules mining (ARM) is an unsupervised learning task. It is used to generate significant and relevant association rules among items in a database. APRIORI and FP-GROWTH are the most popular and used algorithms nowadays for extracting such rules. They are exact methods that consist of two phases. First, frequent itemsets are generated. Then, the latter are used to generate rules. The main drawback of both algorithms is their high execution time. To overcome this drawback, metaheuristics have been proposed. Moreover, to optimize the execution time, since the amount of data is in continuous growth, some parallel architectures can be found in the literature. In this paper, we present an overview of existing literature that investigate the ARM problem using both metaheuristics and parallelism. We will focus on the recent algorithms that tackle these problems using approximate methods and GPU. We will present a non-exhaustive classification of different algorithms according to the type of execution (sequential or parallel) and type of method (exact or approximate). © 2021, Springer. All rights reserved.