Leveraging Geographic Feature Embeddings for Enhanced Location-Based Recommendation Systems
Year: 2024
Type: conference-paper
Abstract: Geographically-aware models are becoming increasingly important in Points of Interest (POI) Recommendation Systems (RSs), particularly with the rise of Location-Based Systems and Social Networks, benefiting various areas and enhancing user experience and engagement. Although current POI RSs are of good quality, they often overlook intrinsic geographic features such as nearby rivers, buildings, and streets in POI’s vicinity, which can significantly influence user preferences. In this study, we propose and evaluate the use of POI type geographic embeddings that incorporate geographic features to enhance POI RSs. The results indicate that this approach improves the effectiveness of POI recommender models.
Authors: Nícolas Moreira Nobre Leite, Claudio E. C. Campelo, Salatiel Dantas Silva
Institution: Universidade Federal de Campina Grande
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4403824108