Analisis Rekomendasi Produk Berdasarkan Segmentasi Pelanggan Menggunakan Algoritma DBSCAN dan FP-Growth

Abstract

Supermarket 212 Mart is a retail business based on sharia cooperatives that has implemented strategy to retain customers. But there is still a drawback, which is just seeing based on the amount of money spent regardless of recency and frequency. In addition, in offering products, 212 Mart does not yet have the right reference. This research combines RFM analysis into data mining techniques to provide more product recommendations good. Parameters considered are customer RFM value, customer segment, and product which are often purchased simultaneously within 1 year of member transaction data using an algorithm DBSCAN and FP-Growth. In customer segmentation, obtained 5 clusters and 31 noise data with an Eps value of 0.060, MinPts 3 and an SI value of 0.4222. Association results using 30% minsup and minconf 70% resulted in cluster 1 having 7 rules, cluster 2 having 6 rules, cluster 3 having 10 rules, cluster 4 has 2 rules, and cluster 5 has 6 rules. The rules formed can be used to direct marketing by recommending products to their respective customers clusters.