An effective explainable food recommendation using deep image clustering and community detection

Mehrdad Rostami, Usman Muhammad, Saman Forouzandeh, Kamal Berahmand, Vahid Farrahi, Mourad Oussalah

Abstract: Highlights•Using a deep learning-based image clustering method, this paper proposes an Explainable Food Recommendation system.•A novel overlapping community detection system is developed and integrated in the recommendation system.•A new similarity score is introduced and incorporated into the recommendation based on a new tendency measure.•A rule-based explainability is introduced to enhance transparency and interpretability of the recommendation outcome.•Experimental results demonstrated the performance of our food recommendation system in comparison to other models.