Market Basket Analysis Using FP-Growth Algorithm On Retail Sales Data

Technological developments make industrial competition increasingly fierce, especially in lifestyle-related industries. In the store where we do our case study, the buyer's interest in buying fashion products has decreased significantly. It has implemented many marketing strategies such as offering discounts, events, or product bundling, but sales still have not reached the target. Therefore, the store needs a system that can predict the pattern of habits and desires of buyers for an item. This system can be based on sales transaction data, which is processed using data mining techniques. In this case, the most suitable method is the association or commonly called market basket analysis. This method works by analyzing the behavior patterns of buyers when buying a product simultaneously at a given time. We choose the FP-Growth algorithm because the process is faster and more efficient. This research output is a system that can count the product association value in transactional data. When we tested 571 transaction data items with minimum support of 4% and a minimum of 50% confidence, four rules were obtained with a lift ratio test value above 1.