Predicting Consumer Purchase Intentions Using Machine Learning

This research investigates the application of machine learning methods in predicting the intention of online shoppers, a very important task in the fast-growing e- commerce industry. In this paper, the performance of some machine learning algorithms, namely K-Nearest Neighbors, Decision Tree, Naive Bayes, Random Forest, Random Tree, Gradient Boosting Tree, and Logistic Regression, is explored for forecasting online purchase behaviors. The features involved in the creation of the dataset for this study are enumerated as the traffic source, session time, amount of product pages visited, and finally, users' feedback. In such a way, the best models that were successful within the framework of the current paper were Gradient Boosting Tree with the rate 88.89% and Decision Tree with the following rate 88.89%, which helped predict their likelihood to buy. These models can help reduce predictive errors and mitigate the variations that exist in consumer behavior-for instance, on personal tastes or browsing behaviors-since they make the targeting of potential buyers better. It also pointed to how machine learning is powerful enough to enhance decision- making in e-commerce by offering more insight into the understanding of consumer intention, thus allowing marketers to shape a better strategy toward increasing sales.