Reasearch on Cross-National E-commerce User Behavior Analysis and Conversion Rate Improvement Based on the Improved XLSTM Algorithm

Abstract The rapid expansion of cross-national e-commerce has brought significant opportunities and challenges in understanding diverse consumer behavior. This study introduces an innovative framework combining the XLSTM (Extended Long Short-Term Memory) model with K-means clustering to analyze user behavior and optimize conversion rates on global e-commerce platforms. XLSTM extends traditional LSTM models by incorporating multi-dimensional cell states, attention mechanisms, and improved memory capabilities, enabling it to effectively capture complex temporal and cross-cultural user behavior patterns. The integration of XLSTM with K-means enhances the clustering process by providing high-quality embeddings that lead to well-defined and stable clusters. Through comprehensive evaluations, the combined approach demonstrates superior performance across key metrics, including Silhouette Score, Davies-Bouldin Index (DBI), and Adjusted Rand Index (ARI), compared to standalone clustering algorithms and traditional LSTM-based methods. Feature importance analysis further identifies coupon usage, visit frequency, and product category interest as the most influential factors in user purchase decisions. The findings highlight the potential of this combined methodology to improve user engagement and optimize marketing strategies for cross-national e-commerce platforms.