A Generative-Augmented Deep Matrix Factorization Model for POI Recommendations

Abstract

Deep matrix factorization has become a pivotal technique for Point of Interest (POI) recommendations, effectively capturing complex user-POI interactions to improve user experiences. However, these systems often face challenges such as data sparsity and the geographical constraints of POIs. To address these limitations, we propose the Geographical Generative-Augmented Deep Matrix Factorization (GeoGMF) model, which integrates deep matrix factorization with a generative augmented network. This model incorporates the learning of user geographical information, including function impacts and transfer costs, while mitigating data sparsity through two novel strategies involving augmented optimization of matrix sampling and matrix update generation. Experiments on real-world datasets demonstrate that our proposed model outperforms existing methods across multiple evaluation metrics.