Online shopping behavior prediction using support vector machines

Abstract

In the recent years, developments with technology and the internet have expanded rapidly. A good majority of business is now done online, and online shopping has become widely used. With the number of people that make use of online shopping, the websites and applications that run these are now able to gather a large amount of data in relation to how consumers behave on their platform. This data can be analyzed to develop a machine learning model that will be capable of predicting consumer behavior in real-time and allow the platform to act accordingly. In this study, a supervised machine learning model, particularly a support vector machine, is developed using an online shopping behavior dataset. Forward sequential feature selection is used with cross-validation in order to determine the most important predictors in the dataset and Bayes' optimization is used with the SVM in order to determine the best set of hyperparameters for the model. With holdout validation, the final accuracy on the test set was found to be 89%.