English Learning Platform Student Behavior Analysis in Buying Courses by Using Data Mining Techniques
Main Article Content
English Learning Platform is an online English learning platform which utilizes subtitled YouTube video as learning platform. This learning platform is said to be personalized, by recommending next material based on how well one performs for each exam given per video. If students are able to learn successfully, they may have a tendency to buy more courses. Our goal is to use data mining techniques to research on the relationship about number of courses that students purchase with their learning behavior. After comparing 5 methods by using 4 machine learning algorithm we found that student score attributes has stronger impact to determine the number of packages that bought by students and the login frequency doesn’t affect the number of purchased courses. The neural network algorithm gives the best accuracy which around 90% compare to another algorithm there are decision tree, random forest, and k-means.