Deep Learning-Based User Information Behavior Mining and Personalized Recommendation Optimization DOI Open Access

Lu Zhang,

Shangxin Chang,

Min Chen

et al.

Journal of Global Information Management, Journal Year: 2025, Volume and Issue: 33(1), P. 1 - 34

Published: March 21, 2025

Personalized recommendation systems have become crucial for enhancing user experience and driving engagement in various online platforms. However, existing methods face challenges accurately modeling high-order user-item relationships, dynamically capturing preferences, effectively utilizing multi-modal data. These limitations hinder their ability to deliver relevant, diverse, context-aware recommendations. To address these challenges, we propose the Graph Attention-based Dynamic Recommendation Framework (GADR). GADR incorporates a graph attention mechanism prioritize relationships dynamically, dual-channel structure simultaneously model long-term short-term unified pipeline integrating textual, visual, behavioral By combining components, ensures adaptability dynamic behavior, improves diversity, enhances ranking quality.

Language: Английский

Deep Learning-Based User Information Behavior Mining and Personalized Recommendation Optimization DOI Open Access

Lu Zhang,

Shangxin Chang,

Min Chen

et al.

Journal of Global Information Management, Journal Year: 2025, Volume and Issue: 33(1), P. 1 - 34

Published: March 21, 2025

Personalized recommendation systems have become crucial for enhancing user experience and driving engagement in various online platforms. However, existing methods face challenges accurately modeling high-order user-item relationships, dynamically capturing preferences, effectively utilizing multi-modal data. These limitations hinder their ability to deliver relevant, diverse, context-aware recommendations. To address these challenges, we propose the Graph Attention-based Dynamic Recommendation Framework (GADR). GADR incorporates a graph attention mechanism prioritize relationships dynamically, dual-channel structure simultaneously model long-term short-term unified pipeline integrating textual, visual, behavioral By combining components, ensures adaptability dynamic behavior, improves diversity, enhances ranking quality.

Language: Английский

Citations

0