Using Random Forest and Support Vector Machine Algorithms to Predict Online Shopper Purchase Intention from E-Commerce Session Data

Abstract

This study explores the use of machine learning algorithms to predict online shopper purchase intention, aiming to provide e-commerce businesses with actionable insights into consumer behavior. The Online Shoppers Purchasing Intention dataset, containing 12,330 session records from an e-commerce site, was analyzed using two classification models: Random Forest and Support Vector Machine (SVM). The models were evaluated based on key performance metrics including accuracy, precision, recall, F1-score, and ROC AUC. Results showed that the Random Forest model outperformed the SVM model, achieving an accuracy of 90.43% and a ROC AUC score of 0.94, indicating strong predictive capability. PageValues and ProductRelated_Duration were identified as the most important features influencing purchasing behavior, with higher values of these features being strongly associated with successful purchases. The study provides valuable insights into the behaviors that drive purchasing decisions in e-commerce, showing that longer engagement with product-related content and higher monetary value pages significantly increase the likelihood of conversion. While the study contributes to understanding online shopper behavior through machine learning, it is limited by the class imbalance in the dataset and the absence of more granular customer data. Future research could address these limitations by incorporating additional features and exploring deep learning models for more accurate predictions. Practical implications of the study suggest that e-commerce businesses can improve conversion rates by optimizing product-related pages and focusing on key user behaviors that are predictive of purchases.