Integrating Semantic Contextualization and Graph Neural Networks for Personalized Content Recommendations in OTT Platforms

Abstract

The exponential growth of Over-the-Top (OTT) platforms has significantly transformed the media consumption landscape, offering users a vast array of content options. However, the challenge of delivering personalized and contextually relevant recommendations persists, primarily due to the complexity and diversity of user preferences and content attributes. This paper presents an innovative framework, the Semantic-Enhanced Graph Model (SEGM), marking a notable advance in tackling these challenges. SEGM integrates semantic embeddings with Graph Neural Networks (GNNs) to capture intricate relationships among users, items, and contextual factors. Our experimental results, validated using the Anime dataset, reveal substantial improvements in recommendation accuracy, relevance, and user engagement. This research not only lays a strong foundation for advancing personalized recommendation systems within dynamic and data-intensive environments such as OTT platforms but also inspires optimism about the potential of this novel approach to shape the future of recommendation systems.