<![CDATA[The development of the retail sector generates a vast amount of transaction data, requiring effective processing methods to understand consumer behavior. This study aims to compare the performance of two popular Association Rule Mining algorithms, namely Apriori and Frequent Pattern Growth (FP-Growth), in extracting purchase patterns using the Market Basket Analysis (MBA) method. The data used is the secondary dataset Market Basket Optimization from Kaggle, which consists of 7,501 transactions. The research method includes stages of data preprocessing, application of one-hot encoding, and implementation of the algorithm using the Mlxtend library in Python with minimum support parameter of 0.01 and minimum confidence of 0.3. The research results show that both algorithms produce an identical number of association rules, namely 63 rules, which validates their accuracy. However, in terms of computational efficiency, FP-Growth shows a significant advantage with an execution time of 0.1547 seconds, approximately 3.7 times faster than Apriori, which takes 0.5731 seconds. The strongest association pattern was found in the relationship between Herb and Pepper and Ground Beef with a lift ratio of 3.29. The conclusion of this study is that FP-Growth is more recommended for large-scale retail data analysis due to its better procedural efficiency without compromising the quality of the generated data.]]>
Copyrights © 2026
