An analysis of customer retention rates by time series data mining

To analyse after-sales services for passenger cars, the traditional research methods only considered the lifelong value of customers when the all services were viewed, while aiming to improve customer retention rates. In this paper, a framework is proposed for time series data mining focused on the time order of a purchase history. The purpose of this study is to acquire the primary reason for a long-term relationship between an automobile dealer and a customer using analyses of relationships between automobile maintenance records and customer retention rates. We attempt to clarify the characteristics of the high retention customers and improve the accuracy of the customer retention predictions as a result of the clustered customer obtained by machine learning. This research is able to show the usefulness of the service science through the application of the engineering methods in the service industry.