Consumer Behavior Analysis Model of Cross-border E-commerce Based on Big Data

The purpose of this study is to build a cross-border e-commerce consumer behavior analysis model based on big data, so as to deeply understand consumers’ preferences, habits and decision-making process in the purchase process, and provide accurate market insight and strategic suggestions for cross-border e-commerce platforms. Through the application of big data technology, this study successfully collected the consumer transaction data of the “The global purchase” platform for nearly one year, and carefully processed and analyzed the data, and extracted key features including purchase frequency, product category preference, payment method selection and so on. In the process of model construction, this study adopts an innovative ensemble learning method, which combines random forest, gradient boosting decision tree (GBDT) and support vector machine (SVM) algorithms, and uses stacking generalization strategy to effectively improve the accuracy of consumer behavior prediction. The empirical analysis results show that the model has high accuracy in predicting consumers’ purchasing behavior and consumption habits. The model analysis shows that electronic products, fashion clothes and household goods are the goods categories with high consumer preference, and most consumers tend to shop on weekends or holidays, and the shopping peak period is from 8: 00 pm to 10: 00 pm. In addition, the model also identifies significant differences among different consumer groups in terms of purchasing preferences and payment methods, which provides a basis for e-commerce platforms to achieve refined market positioning. The consumer behavior analysis model of cross-border e-commerce based on big data constructed in this study not only provides in-depth market insight for cross-border ecommerce platforms, but also provides valuable reference for enterprise inventory management, logistics distribution and marketing strategy formulation.