<![CDATA[Along with the rapid development of technology, the volume and scale of data stored by business entities continue to increase, particularly sales transaction data that contain valuable information to support decision-making and business development. Therefore, this study aims to analyze customer purchasing patterns by combining the Tabu Search and FP-Growth algorithms. Tabu Search is applied as a preprocessing stage to filter and sort transaction data before further analysis using the FP-Growth algorithm as an association analysis method. The results of applying these algorithms are association rules that represent relationships among items and can be used as a basis for business decision-making. The evaluation is conducted using support, confidence, and lift metrics to assess the strength of the generated rules, as well as execution time and the number of itemsets to compare the performance of FP-Growth with and without Tabu Search. The experimental results show that Tabu Search is able to effectively filter itemsets, where at a minimum support value of 0.01 the number of itemsets is reduced from 1,390 to 237, and at a minimum support value of 0.1 from 64 to 34. Although the combination of Tabu Search and FP-Growth requires a longer execution time due to the iterative process of Tabu Search, the resulting patterns are more focused, demonstrating the effectiveness of Tabu Search in improving the efficiency and quality of customer purchasing pattern analysis.]]>
