POI neural-rec model via graph embedding representation

With the booming of Internet Things (IoT) and speedy advancement Location-Based Social Networks (LBSNs), Point-Of-Interest (POI) recommendation has become a vital strategy for supporting people's ability to mine their POIs. However, classical models, such as collaborative filtering, are not effective structuring POI recommendations due sparseness user check-ins. Furthermore, LBSN distinct from other scenarios. respect data, user's check-in record sequence requires rich social geographic information. In this paper, we propose two different neural-network structural deep network Graph embedding Neural-network Recommendation system (SG-NeuRec) Deepwalk on (DG-NeuRec) improve recommendation. combined with representation geographical graph information (called SG-NeuRec DG-NeuRec). Our model naturally combines representations user-POI interaction captures potential interactions under framework neural network. Finally, compare performances these models analyze reasons differences. Results comprehensive experiments real LBSNs datasets indicate performance our model.