Efficient Scalable Recommendation Systems Using Graph-Based Transformers

Priyaranjan Pattnayak

Abstract:

Recommendation systems play a crucial role in personalized content delivery across various industries, including e-commerce, streaming services, and online advertising. Traditional collaborative filtering and deep learning-based approaches struggle to scale effectively to massive datasets while maintaining accuracy. In this paper, we propose a novel recommendation system architecture that leverages Graph-Based Transformers (GBT) to enhance scalability, interpretability, and recommendation precision. By incorporating graph-based relational structures into transformer models, our approach captures complex user-item interactions while maintaining efficiency. Our experimental results on large-scale datasets demonstrate that GBT significantly outperforms baseline methods in terms of precision, recall, and computational efficiency.
Keywords: Recommendation Systems, Transformers, Graph Neural Networks, Collaborative Filtering, Scalability