<![CDATA[The rapid advancement of the digital era has fundamentally transformed the electronic commerce landscape, where the abundance of product options frequently causes information overload that hinders consumer decision-making. This study aims to examine the effectiveness of artificial intelligence implementation in product recommendation systems on e-commerce platforms while comparing the performance of various algorithms employed. The research adopted a comparative experimental quantitative approach using 10,000 transaction records, evaluating three primary algorithms, namely Collaborative Filtering, Content-Based Filtering, and Hybrid Filtering, through Precision, Recall, F1-Score, and RMSE metrics. Findings revealed that AI-based systems achieved an average recommendation relevance rate of 84.6%, substantially surpassing conventional systems at only 51.3%. Among the three algorithms tested, Hybrid Filtering demonstrated the highest performance with an F1-Score of 87.9% and the lowest RMSE of 0.231. The hybrid approach also proved most resilient under cold-start and data sparsity conditions compared to other algorithms. This study concludes that integrating artificial intelligence, particularly through a hybrid algorithm, represents the most optimal strategy for improving product recommendation personalization quality and driving sales conversion on large-scale e-commerce platforms.]]>
Copyrights © 2026
