Perbandingan Kinerja Algoritma Apriori dan Eclat Equivalence Class Transformation Dalam Menerapkan Rekomendasi Barang Diskon pada Data Transaksi

Abstract

This research aims to compare the performance of the Apriori and ECLAT algorithms in applying discounted goods recommendations on sales transaction data. The research method used involves six main stages, namely collecting transaction data from CV SOSO Group Jombang, preprocessing data to clean and prepare data, implementing the Apriori and ECLAT algorithms, forming association rules, evaluating algorithm performance, and visualizing analysis results. The transaction dataset used comes from sales during October and is processed using the Market Basket Analysis approach. Apriori and ECLAT algorithms are tested to find customer purchase patterns by considering the support, confidence, and lift factors in the association rules. Performance evaluation is done by measuring the number of intemsets generated, computation time, and the quality of the rules obtained. The results showed significant differences between the two algorithms. Apriori is easier to implement and provides rules that are more easily understood by decision makers, while ECLAT is superior in time efficiency and memory usage, especially on small to medium-sized datasets. After a comparative analysis, it was found that the ECLAT algorithm is faster in finding transaction patterns with a shorter execution time than Apriori. However, Apriori algorithm is more effective in displaying association rules that are clearer and easier to interpret. With these results, this research recommends the use of algorithms based on specific needs, namely Apriori for more interpretative analysis and ECLAT for analysis with larger and more complex datasets.