Data Mining Implementation Using Frequent Pattern Growth on Transaction Data for Determining Cross-selling and Up-selling (Case Study: Cascara Coffee)

The development and competition of the coffee shop business is increasingly popular nowadays. Cascara Coffee is one of the business shops that must have the right marketing strategy so that the shop's business can survive. This paper presents a study on the use of the FP-Growth association algorithm to process transaction data in order to provide best association parameter in cross-selling and up-selling recommendations for coffee sales. The data used is sales of Cascara Coffee in one year as much as 16,579. Based on the experimental results, the highest lift ratio value obtained is 2,789. In addition, the best association rule is to use a minimum support parameter of $0,20x{10^{ - 2}}$ and a minimum confidence of 0.3. The association rules can be a recommendation for the company to carry out cross-selling and up-selling marketing strategies.