Neural Collaborative Filtering for Restaurant Recommendations: Comparative Evaluation

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Tourism development serves as a crucial catalyst for future economic growth across nations, with dining emerging as one of the foremost leisure travel activities.The growing interest in dining experiences has spurred an increased focus on research related to restaurant recommendation systems (RRS).Deep learning is a powerful technique that can be used to enhance the personalization and scalability of any RRS.This research aimed to build deep learning recommendation systems (DLRS) in the food service industry.We conducted a comparative analysis and an empirical evaluation of multiple recommendation techniques including a matrix factorization approach Singular Value Decomposition (SVD) and deep learning models: Neural Collaborative Filtering (NCF), linear and neat Graph Convolution Network (LightGCN).Performance was measured using standard ranking metrics-Precision, Recall, Mean Average Precision (MAP), and Normalized Discounted Cumulative Gain (nDCG).Among the evaluated approaches, the NCF model achieved the highest values across ranking-based evaluation metrics (e.g., MAP@10 = 0.1404, nDCG@10 = 0.2349, Recall@10=0.4159).While SVD and KNN achieved lower prediction error (RMSE/MAE), NCF demonstrated stronger top-K recommendation performance than the tested baselines.