Next Place Prediction Using GPS traces and Web Search Queries
Ryo Imai (Tokyo Institute of Technology), Kota Tsubouchi, Masamichi Shimosaka (Tokyo Institute of Technology)
18th Annual IEEE International Conference on Pervasive Computing and Communications (PerCom 2020), 2020/3
機械学習 (Machine Learning) データサイエンス (Data Science)
- Thanks to the popularity of GPS-enabled devices, destination prediction, meaning predicting the future location of users, has been investigated as a core technology for various applications in location-based services. For the last decade, predicting daily activities specific to an individual user, such as one's home and office~(familiar destination prediction) is explored, whereas it exhibits difficulties in prediction for unfamiliar destinations such as for shopping on weekends and sightseeing activities. To resolve this limitation, we propose a new framework that exploits web search queries of users as well as GPS. This is inspired by the fact that users tends to perform web searches related to their unfamiliar destinations. To the best of our knowledge, our model is the first attempt that deals with web search queries (connecting natural language processing) and location-oriented research. The experimental results using over 670 users from commercial services show that our proposed method achieves better prediction performance in comparison with the state-of-the-art approaches.