Konsep Inti
Enhancing recommendation systems through knowledge-aware attribute-based representations.
Abstrak
The content discusses the proposal of a method named Knowledge-aware Dual-side Attribute-enhanced Recommendation (KDAR) to improve recommendation systems. It addresses the limitations of existing methods in modeling fine-grained user preferences and leveraging the preference-attribute connection for better performance. The method enhances collaborative filtering based on user and item representations with attribute information from knowledge graphs. Experimental results demonstrate the superiority of KDAR over state-of-the-art baselines across four benchmark datasets.
Directory:
Abstract
Introduces Knowledge-aware Dual-side Attribute-enhanced Recommendation (KDAR).
Introduction
Discusses the importance of knowledge-aware recommendation.
Methodology
Details the components of KDAR: CG & KG Representation Learning, Attribute-based Representation Learning, Multi-level Collaborative Alignment Contrasting, and Model Prediction.
Experiments
Evaluates KDAR's performance against baselines on various metrics.
Related Work
Reviews existing knowledge-aware recommendation methods.
Conclusion, Future Work, and Ethics Statement
Statistik
"Experimental results on four benchmark datasets demonstrate the superiority of KDAR over several state-of-the-art baselines."
"The code of KDAR is released at: https://github.com/TJTP/KDAR."
Kutipan
"User preferences can be modeled using attributes of historical items."
"The connection between user preferences and item attributes can be leveraged to predict users’ interest more precisely."