Analysis of E-learning Customer Data using Data Mining Techniques

The purpose of this study is that provide analysis of e-learning customer data by using machine learning methods include decision tree, Deep belief network (DBN), and support vector machine (SVM). E-learning marketing need to precisely understand their customer in e-learning industry. Deep belief network (DBN) models have been successfully employed to classify problem. This study uses a three-layer deep network of restricted Boltzmann machines (RBMs) to capture the feature of input space of customer data, and after pre training of RBMs using their energy functions, gradient descent training. The customer data of e-learning courses was collected and examined to determine the feasibility of the decision tree, DBN and SVM. This study uses the actual database to select customer's data include "sex", "birth month", "public/private university", "home postal code", and decision variable "classes of study". These customer's datasets are examined through decision trees, support vector machines, and Deep Belief Network Classifier, which provides rules and classifier training results for digital marketing systems. This study can help exploring the relationship of courses, and promote the ability of information for e-learning enterprise. The results show that (1) male students almost selected engineering courses, and (2) female students almost selected business courses. Mainly, except those who live south of Changhua and were born after March or students who were born after September and are "non-Taipei". (3) Students from public or private universities will not affect the students' willingness to study e-learning courses.