Semantic-enhanced neural collaborative filtering models in recommender systems

Article Chemistry, Analytical

Neural Collaborative Filtering with Ontologies for Integrated Recommendation Systems

Rana Alaa El-deen Ahmed et al.

Summary: Machine learning, especially deep learning with neural networks, has achieved remarkable success in various AI problems. The approach of ML differs from classical engineering and ontologies, as it relies on collecting large datasets and processing them through a generic learning algorithm. Combining ontology-based recommendation and ML-based techniques in a hybrid system is a natural and promising method to enhance semantic knowledge with statistical models. This paper presents a novel hybrid recommendation system that blends knowledge-driven recommendations from an ontology with data-driven recommendations generated by classifiers and a neural collaborative filtering. The authors show that the integration of these two worlds provides measurable improvement and enables the transfer of semantic information to ML and statistical knowledge to the ontology. The proposed system also allows for dynamic behavior capturing by updating the ontology with new products and user behaviors.

SENSORS (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Context-Aware Explainable Recommendation Based on Domain Knowledge Graph

Muzamil Hussain Syed et al.

Summary: This paper proposes a context-aware recommender system based on domain knowledge graphs to respond to natural language queries. The system generates candidate recommendations through reasoning over user queries and re-ranks them using contextual information, providing explainable recommendations. Experiments demonstrate that the system performs well in candidate generation and re-ranking.

BIG DATA AND COGNITIVE COMPUTING (2022)

添加到收藏夹

Article Computer Science, Theory & Methods

Ontology based recommender system using social network data

Mohamad Arafeh et al.

Summary: This paper proposes a new framework for sampling Online Social Network (OSN) by using domain knowledge to define tailored strategies, which reduces the budget and time required for mining while increasing recall. The experimental results emphasize the importance of the strategy definition step and the application of ontologies on the knowledge graph in the domain of recommendation analysis.

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE (2021)

添加到收藏夹

Article Computer Science, Theory & Methods