SEGMENTASI PELANGGAN MENGGUNAKAN ALGORITMA FUZZY C-MEANS (FCM) DAN ANALISIS RFM (RECENCY, FREQUENCY, AND MONETERY) PADA DATA PELANGGAN KEDAI

• Main Author: Uus Rusdiana, (Author)
• Format: Book
• Published: 2020-06-18.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upnvj.ac.id/7397/1/ABSTRAK.pdf; http://repository.upnvj.ac.id/7397/13/AWAL.pdf; http://repository.upnvj.ac.id/7397/3/BAB%201.pdf; http://repository.upnvj.ac.id/7397/4/BAB%202.pdf; http://repository.upnvj.ac.id/7397/5/BAB%203.pdf; http://repository.upnvj.ac.id/7397/6/BAB%204.pdf; http://repository.upnvj.ac.id/7397/7/BAB%205.pdf; http://repository.upnvj.ac.id/7397/8/DAFTAR%20PUSTAKA.pdf; http://repository.upnvj.ac.id/7397/9/RIWAYAT%20HIDUP.pdf; http://repository.upnvj.ac.id/7397/14/LAMPIRAN.pdf; http://repository.upnvj.ac.id/7397/12/ARTIKEL%20KI.pdf
• Keywords: QA Mathematics; QA75 Electronic computers. Computer science; QA76 Computer software

Customer segmentation aims to classify customers based on the similarity of characteristics in the form of similar customer interests or demands. To perform customer segmentation, it can be achieved through data mining by implementing clustering techniques. The algorithms that are often used for clustering are the fuzzy c-means algorithm (fuzzy based) and the k-means algorithm (classical based). The selection of the optimal distance metric measurement method for the clustering algorithm is the main problem in this study. Because the use of the distance metric measurement method has a significant effect on the quality of the resulting clusters. The algorithm that will be used in this study is fuzzy c-means clustering with k-means clustering as a comparison which will then try to apply the Euclidean distance, Mahnattan distance, Chebyshev distance, and Minkowski distance measurement methods. The resulting clusters using different distance metrics measurement methods will be evaluated using validity indices including Partition Coefficient Index (PC), Modified Partition Coefficient Index (MPC), and RMSE. The results show that the fuzzy c-means algorithm is superior to k-means with the optimal distance metric, namely Manhattan distance (PC = 0.95, MPC = 0.9, and RMSE = 0.7745) for testing on clusters of 2 and minkowski distance (PC = 0.9338, MPC = 0.9007, and RMSE = 0.8366) for testing on clusters of 3. RFM analysis of the results of customer segmentation using the fuzzy c-means algorithm shows the optimal number of clusters, namely 3 clusters which separate customers into three characteristics, namely high customer retention (value RFM = 3 to 5), moderate customer retention (RFM value = 2 to 3) and low customer retention (RFM value = 1).