Temporal Preference and Knowledge-Aware Collaborative Attentive Network for Electrical Material Recommendation
1月 1, 2025·,,,,,,·
0 分钟阅读时长
Jie Shen
Lei Chen
Guixiang Zhu
Jie Cao
Weiping Qin
Yihan Chen
Yiheng Lu
摘要
In the field of electrical power material management, it is paramount that
users receive accurate recommendations regarding the electrical power materials
they require. In recent years, increasing attention has been given to Knowledge
Graph-based recommender systems for their capacity to effectively tackle cold start
and data sparsity challenges. However, many of these systems still lack the capability
to consider users’ historical behavioral patterns and model dynamic changes in user
interests. To remedy this gap, the paper proposes an Electrical Material Recommendation
model based on Temporal Preference and Knowledge aware collaborative attentive network,
named EMR-TPK. EMR-TPK establishes associations between users and materials through
knowledge graph embedding along with an attention mechanism. It utilizes a temporal
dynamic attention network to characterize users’ temporal interests and historical
behavioral patterns. By incorporating these patterns into the model, more accurate
material recommendations can be generated. We conducted extensive experiments across
three datasets: electric power materials, music, and books. Our results indicate
that the EMR-TPK model exceeds advanced methods in recommendation accuracy, with
AUC improvements of 0.36% for electric power materials, 0.95% for music, and 2.54%
for books.
类型
出版物
Web Information Systems Engineering – WISE 2024