Mining and Exploration of Credit Cards Data in UAE Sarween Zaza The British University in Dubai Dubai, UAE 120155@student.buid.ac.ae Mostafa Al-Emran Al Buraimi University College Al Buraimi, Oman malemran@buc.edu.om Abstract— Credit cards have become an essential element in the banking industry. Credit cards add a significant value for the banks. Mining credit cards can find interesting patterns among different variables that may be used in the future by the policy makers for building their future policy. In this study, we have investigated the credit card-holder’s behavior in order to predict the market segmentation. An online questionnaire survey regarding credit card usage has been used for data collection. Two techniques have been applied on the collected data, Decision Trees and K-means through the use of training and testing sets. Results indicated how people are grouped based on their income which in turn will help in building the appropriate decision on which region needs to be targeted. Moreover, results revealed different work sectors for the credit card-holders and which type of credit card is used with regard to their income. Keywords—Credit Cards; Data Mining; UAE. I. INTRODUCTION Credit cards are considered as an important product for industrial banking. Lots of judgmental techniques drive the decision-making process. The business framework is usually built based on the bank database’s credit card. Commercial banks began concentrating on expanding the card business. As a result, it has been observed that China started to exceed the U.S. as the biggest world’s market for the total credit card numbers by the year 2020 Based on SHANGHAI (Dow Jones) MasterCard Incorporation [1]. UAE as a one of the growing commercial countries starts with a limited targeting of this market. Depending on the conducted survey in UAE, this study attempts to support the driving of decision-making in banks in order to develop marketing segmentation. The marketing is applied by adopting strategies and building new plans. Market knowledge helps bankers to achieve their significant objectives in their promotions. Accordingly, this will increase their sales and leverage their customers’ satisfaction. Based on the conducted credit card survey data in the UAE, this paper has selected some related indicators by taking into account customer income. Decision trees and K-means techniques have been used in order to choose the best prediction technique in order to supply feedback to stockholders. Results indicated that banks can plan and adopt market strategies. It is helpful to segment customers related to specific indicators. Moreover, application of this technique will improve the bank product offerings when compared with the developed nations according to the UAE vision 2030. II. LITERATURE REVIEW Data mining refers to elicitation, a valuable pattern or principles from a spatial database. The process of data mining is used to distinguish a particular issue or issues. Data mining recalls the necessary database and applies such techniques to test the specified data with a final goal of achieving an outcome used to drive the decisions on strategic marketing [2]. Data mining is a Knowledge Discovery application in Databases (KDD). Data mining techniques can be applied either as perspective, prospective or detection modeling [3]. Partition objects are applied to clusters or segments with a significant database with the same characteristics. This kind of segmentation is beneficial in targeting markets. Partitioning has been used in different domains such as insurance, finance, marketing, telecommunication, banking or any other fields that provides the user with a breakdown prediction. Data mining refers to the detection of some useful list of credit cards types, and it is a powerful tool in providing individuated service of credit cards in banks. A study by [4] characterized the implementation of data mining in credit cards business comprehensively. Data mining is categorized into two groups [1]. The first group represents the market and customer growth, while the second one represents the risk control containment modeling and scoring of the implementation model. The use of such mined data enables the selection of targeted customers and analyzing a data- mining model that could be applied to the granting and adopting of existent credit card lines. Furthermore, the use of such mined data helps in the preservation or retention of customers within the client group and also allows for informed upgrades to customer loyalty. A study by [5] partitions the customers into three groups with regard to their contributions such as high quality, star and common. The insight provided with mined data allows for explicitly considering the customers who are always changing, presents the needs to look for new customers and offer new services for the inherent customers. Different research papers have been reviewed in order to discuss mining of credit cards data. All data is based upon an actual data set from banks based in China, USA and Egypt. 2015 Fifth International Conference on e-Learning 978-1-4673-9431-4/15 $31.00 © 2015 IEEE DOI 10.1109/ECONF.2015.57 267 2015 Fifth International Conference on e-Learning 978-1-4673-9431-4/15 $31.00 © 2015 IEEE DOI 10.1109/ECONF.2015.57 275 2015 Fifth International Conference on e-Learning 978-1-4673-9431-4/15 $31.00 © 2015 IEEE DOI 10.1109/ECONF.2015.57 275 2015 Fifth International Conference on e-Learning 978-1-4673-9431-4/15 $31.00 © 2015 IEEE DOI 10.1109/ECONF.2015.57 275