PERBANDINGAN KINERJA ALGORITMA APRIORI DAN EQUIVALENCE CLASS TRANSFORMATION (ECLAT) DALAM MENEMUKAN POLA PEMBELIAN PADA DATA TRANSAKSI MINIMARKET

Abstract

This study compares the performance of the Apriori and ECLAT algorithms in analyzing sales transaction data from a minimarket. The research focuses on examining both algorithms' efficiency in terms of execution time and memory usage when identifying frequent itemsets and generating association rules. Given the limited variety of products sold in a minimarket, a lower minimum support (0.001) and minimum confidence (0.005) were applied to ensure meaningful results, as higher thresholds resulted in no significant findings. The first test evaluated the time required to find frequent itemsets, revealing that ECLAT consistently outperformed Apriori with an average execution time of 0.71634 seconds compared to Apriori's 4.88256 seconds. The second test assessed the time taken to generate association rules, where ECLAT again showed slightly better performance, averaging 0.01352 seconds versus Apriori's 0.01618 seconds. Memory usage tests showed that ECLAT was more efficient, using an average of 0.12436 MB to find frequent itemsets and 0.01052 MB to generate association rules, compared to Apriori's 0.1385 MB and 0.01136 MB, respectively. The results indicate that the ECLAT algorithm is generally more effective for analyzing sales transactions in a minimarket environment, particularly when handling large datasets and when computational efficiency is critical. The findings provide valuable insights for selecting the appropriate algorithm to optimize marketing strategies and inventory management in retail settings.

Keywords: Market Basket Analysis, Apriori, Assocation Rule, ECLAT