Comparison of Two Methods of Association Rules Based on Running Time for Different Number of Transactions and Support Values

Association rule mining is a widely-used data mining method aimed at discovering the hidden patterns and connections within collections of transactional data, usually interpreted as so-called “if-then” statements. This paper aims to provide an overview of the well-known algorithms in this area, compare their advantages and disadvantages, and then apply them to a specific dataset. Moreover, we focused on the performance analysis of the two implemented algorithms based on running time for different number of transactions and support values. We found that the incorrect choice of parameters in the search for association rules can have two serious consequences: an increased number of rules or extended computing time. The results can help select the optimal algorithm for a specific task.