Influence of Minimum Support on the Performance of the Apriori Algorithm

Understanding consumer behavior patterns is very important to consider to increase sales. The method often used to analyze purchasing patterns made by consumers is the association method or association rule mining. The Apriori algorithm is one of the most commonly used algorithms in data mining to find high-frequency patterns in transaction data. The main challenge of the Apriori algorithm is related to the issue of computational efficiency, namely the problem of storage space efficiency and execution time efficiency. The minimum support and confidence values determine the formation of association rule patterns using the Apriori algorithm. In several studies, association patterns were built using different minimum support values. This study aims to analyze the effect of minimum support values on the performance of the Apriori algorithm. The dataset used in this study is retail transaction Groceries Data. Based on the results of the experiments conducted, it was obtained that each increase in the minimum support value of 0.0025, resulted in a decrease in the average number of frequent item-sets, execution time requirements, and memory requirements of $19.05 \%, 28.12 \%$, and $15.73 \%$, respectively.