Forecasting e-learning Course Purchases Using Deep Learning Based on Customer Retention

Retention analysis is essential from the point of view of customer behavior, including purchases of products. Existing purchasing forecasting approaches are based on analyzing many user profile characteristics, which are only sometimes available to e-learning course providers. The aim of this research is to develop a method for forecasting e-learning course purchases using deep learning based on customer retention. The multi-layer perceptron and long short-term memory are used. Analyzing the research results, it can be seen that the long short-term memory network achieved significantly better results than the multi-layer perceptron. Using a developed approach, there is no need to acquire additional data characterizing the user, such as demographic data. From the business point of view, it is a crucial benefit for new customers who are not registered in the customer relationship management system, but use anonymously this system via the website.