چكيده به لاتين
Abstract:
Today, disregarding customer communication management and marketing activities in electronic banking may cause customers to turn away and leave. According to their important role in electronic banking services, customers have become interested in POS. Lack of satisfaction in the quality of POS services, mostly ensued by the failure of transactions, can make customers reject using banks` POS.
This study aims to use data mining association rules to interpret customer clusters of the sales terminal of Saderat bank in different regions of Tehran. This study uses the K-Means clustering model to cluster customer transactions. Accordingly, customer transactions are divided into three parts. In each generated part, after studying its transactions` features, decision tree and classification algorithms of Rough set theory are used to predict the success or failure of the POS transactions. The Rough set theory provides one of the most powerful types of association rules, particularly in the presence of uncertainty. Subsequently, models are compared in terms of their accuracy. Results indicate that due to considering the uncertainty of rules, the Rough set theory predicts the solution variable (operation result) more accurately than the decision tree for all clusters.
Keywords: Transaction, POS, Association Rules, Decision Trees, Rough set theory, Saderat Bank.
