Citation Context

...he medium e-banking customers, both individuals and companies, were used. A Boolean value is assigned to each different payment type depending on whether the payment has been conducted by the users or not. A sample of the above data set is shown in Table 2. User SWIFT P.O. FUNDS TRANSFER FORWARD FUNDS TRANSFER GTOPPC S.O. SII P.O. VAT P.O. CREDIT CARD P.O. … … … … … … … … User103 F T T F F F T User104 T T T F T T F User105 F T T T T T T User106 F T F T F F T … … … … … … … … Table 2 In total the sample contains 298 medium e-banking customers. In order to discover association rules the Apriori [SPSS (2002)] method was used. These rules are statements in the form if antecedent then consequent. 4. Experimental results 7 As seen in the histogram of Figure 2, RFM distribution is high over values less than 1.000.This is a natural trend since, as concluded in paragraph 1, 80% of the customer exhibits low RFM Factor. Figure 2 Application of the K-means algorithm results in the 4 clusters of Figure 3. Next to each cluster one can see the number of appearances as well as the average value of each variable. Figure 3 8 The above clustering results in the distribution of Figure 4. The similarity of this di...

Citation Context

...d by the use of the model follow: Decision support and Decision making. Future revenue forecast. Customer profitability. Predictions concerning the alteration of customers’ position in the pyramid. Understanding the reasons of these alterations. Conservation of the most important customers. Stimulation of inactive customers. Στο 16ο Εθνικό Συνέδριο Ελληνικής Εταιρίας Επιχειρησιακών Ερευνών, 2003 2 Figure 1 Essentially RFM analysis suggests that the customer exhibiting high RFM score should normally conduct more transactions and result in higher profit for the bank. RFM analysis [SPSS (2001), Madeira, S.A. (2002), COMPAQ (2001), Im, K. and Park, S. (1999)] nowadays can be conducted by the use of Data Mining methods like clustering. These methods contribute to the more efficient determination and exploitation of RFM analysis results. Determination of association rules concerning bank data is a challenging though demanding task since: • The volume of bank data is enormous. Therefore the data should be adequately prepared by the data miner before the final step of the application of the method. • The objective must be clearly set from the beginning. In many cases not having a clear ...