Applying the Neural Network Method in Analyzing the Preferences of Food Store Customers

Abstract

Abstract: The primary goal of the present study is to use the neural network method in analyzing the preferences of customers of food chain stores. Valuable customers were identified using neural network methods and clustering methods. Then, valuable customers were classified by classification methods. Then, the preferences of valuable customers were identified. It was done by the association rules approach.  Additionally, using classification and clustering methods, useful patterns were found to identify and analyze the behavior of outbound and non-outbound customers. After a series of data preparation and pre-processing operations, each customer's information and their transactions were determined.  The primary data were based on each customer's transaction. After the data preparation and pre-processing operations, a set of data was obtained that was related to each customer and recorded their information.  Neural network methods were used to classify valuable customers and leaving customers and this network was compared with other methods. Concerning K-means clustering, the output of this algorithm could identify valuable customers and analyze the customers who left or were loyal.  Association rules were also used for the cluster of valuable customers to identify the preferences of valuable and non-leaving customers.