The Best Association Model on Online Retail Datasets

The online retail industry is rapidly expanding, necessitating effective analysis of sales transaction data to leverage market opportunities. This study investigates the challenges of analyzing sales transaction data from the Online Retail dataset by comparing the performance of FP-Growth and Apriori algorithms. The primary objectives include evaluating the most efficient association algorithm, identifying effective association models in complex datasets, and assessing the algorithms' efficiency in processing data. The methodology involves analyzing association data using FP-Growth and Apriori, with lift metrics for measuring association strength and processing time for efficiency evaluation. The results indicate that both algorithms produced association models with a highest support value of 0.262 and confidence of 0.919. Despite generating similar frequent itemsets, the Apriori algorithm's highest support value appeared at index 83, whereas FP-Growth's was at index 149. Notably, FP-Growth demonstrated superior processing speed, completing operations in 0.216 seconds compared to Apriori's 0.227 seconds. This study concludes that while both algorithms are effective, FP-Growth is preferable for faster processing in large datasets, enhancing the efficiency of sales transaction analysis in online retail.