Consumer Purchase Behavior Prediction on E-commerce Platforms Based on Machine Learning Fusion Algorithm

Abstract

To enhance the precision of predicting consumer purchasing behavior, the author conducts a study focused on forecasting buying patterns on e-commerce platforms through the use of machine learning fusion techniques. The research specifically integrates logistic regression and support vector machine algorithms to analyze shopping behavior data from Alibaba’s e-commerce platform. The experiment revealed that, out of 1,445 test samples fed into the model, 571 were predicted to exhibit purchasing activity on the 32nd day, as indicated by a prediction outcome of ”1.” Compared with the samples with actual purchasing behavior on the 32nd day, their F1 score was 7.77%. The practical results show that the fused model is more accurate in prediction performance than a single algorithm model.