ANALISIS ALGORITMA FP-GROWTH DAN APRIORI UNTUK MENEMUKAN MODEL ASOSIASI TERBAIK PADA DATASET ONLINE RETAIL

Abstract

In the digital era, the online retail industry is growing rapidly and is becoming an important sector. However, challenges arise in the analysis of sales transaction data on the Online Retail dataset. This study aims to overcome problems in the analysis of sales transaction data in the Online Retail dataset. The main focus includes selecting the optimal association algorithm between FP-Growth and Apriori, identifying relevant association models on complex datasets, and the efficiency and performance of algorithms in processing sales transaction data. The method used is association data processing using the FP-Growth and Apriori algorithms. Implementation of the association rule involves adding a lift metric as a measure of association strength. Measurement of processing time is also carried out to determine the efficiency of implementation. The results showed that FP-Growth and Apriori could produce an association model with the same frequent itemset and value matrix, namely a support value of 0.12 and a confidence value of 0.96, but there were differences in the resulting model order. The Apriori algorithm produces a model with the highest support value at index 18, while FP-Growth at index 10. In addition, the FP-Growth algorithm shows an advantage in faster processing time (0.004 seconds) compared to Apriori (0.007 seconds). This research provides a better understanding of the use of association algorithms in the context of the online retail industry.