Location recommendation by combining geographical, categorical, and social preferences with location popularity
Yaxue Ma, Jin Mao, Gang Li · Proceedings of the Association for Information Science and Technology · 2019
ABSTRACT Location recommendation could not only increase location visibility, but free users from information overload dilemma. We propose a novel location recommendation method, which first exploits geographical proximity of locations and check‐in history of users' friends to measure geographical preference and social preference respectively, and innovatively estimates categorical preference by calculating semantic similarity between location tags based on category hierarchy. Then, all of three preferences are integrated, and location popularity is introduced as a coefficient to determine the priority of locations to be recommended. The preliminary result on a publicly available real‐word dataset shows that the proposed method significantly outperforms two state‐of‐the‐art baselines. This study expands the methods of location correlation measurement, and can meet user's personalized information needs.