Cross-Platform E-Commerce Product Categorization and Recategorization: A Multimodal Hierarchical Classification Approach

This study addresses critical industrial challenges in e-commerce product categorization, namely platform heterogeneity and the structural limitations of existing taxonomies, by developing and deploying a multimodal hierarchical classification framework. Results show that CLIP multimodal embeddings combined via an MLP-based late-fusion strategy achieve the highest hierarchical F1 (98.59%), outperforming unimodal baselines. To address shallow or inconsistent categories, we further introduce a self-supervised “product recategorization” pipeline using SimCLR, UMAP, and cascade clustering, discovering new, fine-grained categories with cluster purities above 86%. Cross-platform experiments reveal a deployment-relevant trade-off, complex late-fusion methods maximize accuracy with diverse training data, while simpler early-fusion methods generalize more effectively to unseen platforms. Finally, we successfully demonstrate the framework’s industrial scalability through deployment in EURWEB’s commercial transaction intelligence platform.

2025 IEEE International Conference on Big Data (BigData), Macau, China, pp. 2549-2554