Pre-training Context and Time Aware Location Embeddings from Spatial-Temporal Trajectories for User Next Location Prediction

Authors

  • Yan Lin School of Computer and Information Technology, Beijing Jiaotong University, Beijing, China Beijing Key Laboratory of Traffic Data Analysis and Mining, Beijing, China
  • Huaiyu Wan School of Computer and Information Technology, Beijing Jiaotong University, Beijing, China Beijing Key Laboratory of Traffic Data Analysis and Mining, Beijing, China CAAC Key Laboratory of Intelligent Passenger Service of Civil Aviation, Beijing, China
  • Shengnan Guo School of Computer and Information Technology, Beijing Jiaotong University, Beijing, China Beijing Key Laboratory of Traffic Data Analysis and Mining, Beijing, China Key Laboratory of Transport Industry of Big Data Appalication Technologies for Comprehensive Transport, Beijing, China
  • Youfang Lin School of Computer and Information Technology, Beijing Jiaotong University, Beijing, China Beijing Key Laboratory of Traffic Data Analysis and Mining, Beijing, China CAAC Key Laboratory of Intelligent Passenger Service of Civil Aviation, Beijing, China Key Laboratory of Transport Industry of Big Data Appalication Technologies for Comprehensive Transport, Beijing, China

DOI:

https://doi.org/10.1609/aaai.v35i5.16548

Keywords:

Mining of Spatial, Temporal or Spatio-Temporal Da

Abstract

Pre-training location embeddings from spatial-temporal trajectories is a fundamental procedure and very beneficial for user next location prediction. In the real world, a location usually has variable functionalities under different contextual environments. If the exact functions of a location in the trajectory can be reflected in its embedding, the accuracy of user next location prediction should be improved. Yet, existing location embeddings pre-trained on trajectories are mostly based on distributed word representations, which mix a location's various functionalities into one latent representation vector. To address this problem, we propose a Context and Time aware Location Embedding (CTLE) model, which calculates a location's representation vector with consideration of its specific contextual neighbors in trajectories. In this way, the multi-functional properties of locations can be properly tackled. Furthermore, in order to incorporate temporal information in trajectories into location embeddings, we propose a subtle temporal encoding module and a novel pre-training objective, which further improve the quality of location embeddings. We evaluate our proposed model on two real-world mobile user trajectory datasets. The experimental results demonstrate that, compared with the existing embedding methods, our CTLE model can pre-train higher quality location embeddings and significantly improve the performance of downstream user location prediction models.

Downloads

Published

2021-05-18

How to Cite

Lin, Y., Wan, H., Guo, S., & Lin, Y. (2021). Pre-training Context and Time Aware Location Embeddings from Spatial-Temporal Trajectories for User Next Location Prediction. Proceedings of the AAAI Conference on Artificial Intelligence, 35(5), 4241-4248. https://doi.org/10.1609/aaai.v35i5.16548

Issue

Section

AAAI Technical Track on Data Mining and Knowledge Management