Abstract
The exponential growth of cross-border e-commerce has
fundamentally transformed global retail dynamics, introducing
complex challenges in supply chain management, demand
forecasting, and personalized marketing. As consumer behaviors
diversify across cultural and geographical boundaries, traditional
analytical methods often fail to capture the nuanced patterns
inherent in multi-jurisdictional transaction data. This paper
proposes a comprehensive predictive analytics framework that
integrates big data clustering techniques with association rule
mining to identify latent market trends and optimize decision
making processes. By leveraging K-means clustering, we segment
heterogeneous customer bases into distinct behavioral profiles,
reducing the dimensionality of the dataset while preserving critical
variances. Subsequently, we apply the Apriori algorithm to uncover
strong association rules within these clusters, revealing purchasing
interdependencies that inform predictive inventory positioning and
targeted promotional strategies. The study utilizes a massive dataset
comprising transactional records from diverse international
markets to validate the model. Our findings demonstrate that this
hybrid approach significantly enhances the accuracy of trend
prediction compared to standalone methods, providing actionable
insights for reducing logistics costs and improving customer
satisfaction in the volatile cross-border landscape.

This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright (c) 2026 Hiroshi Tanaka , Eleanor Sterling (Author)