A Comparative Study of Different Data Mining Algorithms with Different Oversampling Techniques in Predicting Online Shopper Behavior

Abstract

Nowadays buying over the internet has become very popular among online users. This action or attitude has increased due to the facilities and features that e-commerce sites offer to users, such as availability, accessibility, no crowds, easy price comparisons, etc. However, the number of actual buyers is still very low compared to the number of total visitors of these sites. Therefore, this paper will study the behavior of online shoppers to predict whether they will buy a product or not. The study first compares several classification algorithms against each other, then tries to enhance the results using different oversampling techniques. Results show that the RF algorithm achieved the best results and regarding oversampling, the SVMSMOTE exceeded the other methods.