Customer Behavior Analysis by Association Rules Mining

Abstract

Emerging technology and the accompanying globalization further strengthen the conditions of competition. At this point, the concept of “difference making-innovation” has become one of the important concepts in today's world. Diversifying businesses can quickly adapt to this development process and take up its place in the development / growth cycle. To making a difference in terms of businesses means to be able to adapt quickly to developing management understanding and to reflect this difference in their business processes. In this paper, we are looking for an answer to two important questions for a retail firm that serves in textile sector in İstanbul/Turkey. Our retail firm has 56 different sales points throughout Istanbul. 18 of them are in the Anatolia and 38 of them are in the European side of İstanbul. The firm produces and sells textile products with different price ranges and different specifications. According to the sales habits of the customers, what products should be sold in the product range of the company and which products should be sold in two branches that we decided? In our study, we have proposed a model that explains customer behavior and makes forward-looking forecasts with the Data Mining application used in Customer Relationship Management to find out the answers to these questions. The model was evaluated by Association Rules Mining (ARM) based on past sales slips, the results were obtained and future estimates were made. The purpose in the association rule; to find the relationships between the products purchased by customers during shopping, and to determine the buying habits of customers in line with this relationship data. Also in this paper the tourism effects are tried to be examined because there are two different branches. One of them has national customers but the other one has more foreign customers