Deep-Distributed-Learning-Based POI Recommendation Under Mobile-Edge Networks

With the rapid development of edge intelligence in wireless communication networks, mobile-edge networks (MENs) have been broadly discussed academia. Supported by considerable geographical data acquisition ability mobile Internet Things (IoT), MENs can also provide spatial locations-based social service to users. Therefore, suggesting reasonable points-of-interest (POIs) users is essential improve user experience MENs. As simple user-location usually sparse and not informative, existing literature attempted extend feature space from two perspectives: 1) contextual patterns 2) semantic patterns. However, previous approaches mainly focused on internal features users, yet ignoring latent external among them. To address this challenge, article, a deep distributed-learning-based POI recommendation (Deep-PR) method proposed for situations In particular, hidden components both local global subspaces are deeply abstracted via representative learning schemes. Besides, propagation operations embedded iteratively reoptimize expressions space. The successive effect above aspects contributes lot more fine-grained spaces, so that accuracy be ensured. Two types experiments carried out three real-world sets assess efficiency stability Deep-PR. Compared with seven typical baselines respect four evaluation metrics, obtained results overall performance Deep-PR excellent.