Data-Driven Customer Online Shopping Behavior Analysis and Personalized Marketing Strategy

Abstract

In today's highly competitive market environment, personalized marketing has become an important means for enterprises to gain competitive advantages. In order to better meet customer needs, companies need to accurately identify and classify customers to implement more refined market strategies. This study focuses on the customer classification problem. Based on several classic deep learning models, the BiLSTM-TabNet model is designed, and the Whale Optimization Algorithm (WOA) is introduced to further improve the model performance, thereby improving classification accuracy and practicality. Experimental results show that this model has achieved excellent performance on each data set, has higher accuracy and AUC value than the baseline method, and has advantages over other control models in comparative experiments. This research provides solid support for the implementation of personalized marketing strategies.