A hybrid recommender system based on data enrichment on the ontology modelling

Area of Science:

• Computer Science
• Artificial Intelligence
• Information Retrieval

Background:

• Recommender systems predict user preferences using user behavior data.
• Hybrid model-based systems require pre-trained models for recommendations.
• Ontologies provide semantic information and relationships for data modeling.

Purpose of the Study:

• To enhance matrix factorization model accuracy in recommender systems.
• To enrich user-item matrix information using ontologies.
• To mitigate the cold start problem in model-based recommender systems.

Main Methods:

• Utilized ontology to enrich the user-item matrix.
• Integrated item-based and user-based collaborative filtering techniques.
• Combined semantic similarity from ontology with rating patterns to reduce data sparsity.

Main Results:

• Reduced data sparsity from 0.9542% to 0.8435%.
• Achieved a lower Root Mean Square Error (RMSE) of 0.9298 compared to baseline (0.9642) and existing methods (0.9492).
• Demonstrated improved accuracy on the MovieLens 100k dataset.

Conclusions:

• The proposed method enhances recommender system datasets by integrating collaborative filtering techniques.
• The approach effectively reduces data sparsity and improves recommendation accuracy.
• The system outperforms baseline and existing methods in accuracy and sparsity reduction.