Citation

Liau, Kai Ze and Haw, Su Cheng and Naveen, Palanichamy and Jayaram, Jayapradha (2026) Comparative Evaluation of Generative AI Models for e-Commerce Product Recommendation. In: 9th International Conference on Artificial Intelligence and Big Data, ICAIBD 2026, 29 May 2026 - 31 May 2026, Chengdu, China.

Text 10.pdf - Published Version Restricted to Repository staff only Download (639kB)

Abstract

Recommender systems are of significant importance to e-commerce platforms, aiming to assist users in finding items suited to their interests from the large catalogue of available products. Although there has been much research into generative AI models, direct empirical comparisons among specific models on e-commerce datasets remain limited. This paper specifically empirically compares these models on e-commerce based datasets. In this work, we perform such experiments using three example models; RecVAE, DiffRec and BERT4Rec on two Amazon ecommerce product review datasets namely Magazine Subscriptions and Gift Cards. Performance is evaluated using seven metrics: Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Recall, Mean Reciprocal Rank (MRR), Normalised Discounted Cumulative Gain (NDCG), Hit Rate (HR), and Precision at top-k where k=10 and k=20. Results show that RecVAE provides the best ranking results overall, DiffRec produces almost error-free predictions and BERT4Rec is most sensitive to choice of hyperparameter settings as its performance varies significantly with different hyperparameter configurations.The findings provide practical insights into model selection for e-commerce recommender system.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	Recommender systems , variational autoencoder
Subjects:	H Social Sciences > HF Commerce > HF5001-6182 Business > HF5546-5548.6 Office management > HF5548.32-.34 Electronic commerce
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	01 Oct 2026 07:14
Last Modified:	01 Oct 2026 07:14
URII:	http://shdl.mmu.edu.my/id/eprint/16789

Downloads

Edit (login required)