Applications of data mining algorithms for customer recommendations in retail marketing

Abstract

In recent years, researchers have highlighted how large volumes of data can be transformed into information to determine customer behaviors, and data mining applications have become a major trend. It has become critical for organizations to use a tool for understanding the relationships between data to protect their marketplace by increasing customer loyalty. Thanks to data mining applications, data can be processed and transformed into information, and in this way, target audiences can be determined while developing marketing strategies. This chapter aims to increase the market share with products specific to the customer portfolio, introduce strategic marketing tools for retaining old customers, introduce effective methods for acquiring new customers, and increase the retail sales chart, based on purchasing habits of customers. The data set was collected under pandemic conditions during the COVID-19 process and analyzed to support retail businesses in their online shopping orientation. By examining the local customer base, it was assumed that the customer group would display similar behaviors in online or teleordering methods, customer identification and order estimation were made to follow an effective sales policy. Segmentation was performed with data mining applications, and the grouped data were separated according to their similarities. The data set consisting of demographic characteristics and various product information of the enterprise's customers were analyzed with Decision Tree and Random Forest, which are data mining methods, the best performing algorithm in the data set was selected by comparing the performance of the methods. As a result of the findings, appropriate suggestions were given to the business to determine the purchasing tendencies of the customers and to increase the level of effectiveness in sales-marketing strategies. In this way, materials were presented to assist the enterprise in developing strategies to increase the number of sales by taking faster and more accurate action by avoiding the time and expense that would be lost by the trial-error method.