| Summary: | 碩士 === 元智大學 === 資訊管理學系 === 105 === With the popularity of e-commerce and changes in people's consumption habits, there are now more and more people choose to shop online, rather than to the traditional department stores or retail stores to buy the required items. The coming of e-commerce have left a powerful impact on sales and visitor numbers of traditional department stores, in this paper, we want to help them face the impact with database. We use data mining approaches to find pattern as customer value through basic customer and transactional data. With RFM model the most famous model in customer relationship management and the age, gender in the basic customer data as the basis for mining, we calculate the value of each customer.
Finally, we use customer value to list direct mailing lists, predict whether the customer will churn next year and predict whether the customer will repurchase same period next year. In this paper, we want to improve the performance of direct mailing by using different sort of customer value. In customer churn and repurchase prediction, we use customer value as training data, through training the classifier to help us predict customer churn and repurchase.
|
|---|
