DATA MINING USING FUZZY METHOD FOR CUSTOMER RELATIONSHIP MANAGEMENT IN RETAIL INDUSTRY
Abstract: A problem that
appears in a retail industry with a great quantity of customers is how to
identify potential customers. A retail industry could identify their best
customer through customer segmentation by applying data miningand customer
relationship managementconcept. This paper presents data mining process from
customer's data in retail company by combining fuzzy RFM model with fuzzy
c-meansand fuzzy subtractive algorithm. The dataconsisted of 3.000.000 rows of
transaction data from 2006 to 2010. The data transferred to 499 RFM data for
each time period selected. Experiments tried two to six clusters by changing
the value of cluster number (FCM) and radii(fuzzy subtractive). The clustering
result will then be classified to determine customer segmentation using fuzzy
RFM models. The modified partition coefficient and partition entropy indexes
used to evaluate the performance of both clustering algorithm.The results
indicate that FCM has a higher validity rate than fuzzy subtractive. Fuzzy RFM
segmentationindicates that fuzzy subtractive can not form a cluster that are
categorized as potential customers, therefore FCM is more appropriate for
customer segmentation in retail industry.
Keywords: fuzzy RFM
model,fuzzy c-means, fuzzy subtractive, modified partition coefficient,
partition entropy
Penulis: Yohana Nugraheni
Kode Jurnal: jptkomputerdd130236