The Application of Machine Learning in Online Purchasing Intention Prediction

Abstract

With the rapid development of computer technology and e-commerce, the consumption concept of consumers has undergone great changes. Online shopping has gained increasing popularity. For online shopping platforms like Amazon, a reliable analysis system of online purchasing behaviors enables them to have a better understanding of consumer psychology, and then work out better business strategies to increase sales. This paper focus on factors that may influence visitors’ shopping intention based on Online Shoppers Purchasing Intention Data Set, which consists of aggregated page view data during the visit session along with other user information. By descriptive statistical analysis, we find that variables like time spend and page values are positively correlated with shopping intention, while variables including bounce rate and exit rate are negatively correlated with shopping intention. Three online purchasing intention prediction models are constructed based on corresponding classical machine learning models (Logistic Regression, Decision Tree and Random Forest). As a result, random forest performs best in Online Shoppers Purchasing Intention prediction, with train accuracy 89.5% and test accuracy 87.5% respectively.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 13 citations exist upstream.