With the popularity of internet lifestyle, various of mobile software are widely used to manage and record daily lives. However, providing personalized functions for travel data management, statistics and display for users is difficult. In this paper, we analyze the problems in the current personalized Point-of-interest (POI) recommendation algorithm, and on this basis, introduce a new neural network model. Our model applies the encoding-decoding framework based on GRU (Gate Recurrent Units). Besides, different attention mechanisms are also applied for considering the impact of contextual factors and check-in sequences for model prediction performance, including the temporal and multi-level context attention mechanisms. We have conducted two experiments with different check-in datasets. The result provided the better prediction performance of our model than several latest models.
Temporal and Multi-level Context Attentive GRU Neural Networks
Lect. Notes Electrical Eng.
International Workshop of Advanced Manufacturing and Automation ; 2020 ; Zhanjiang, China October 12, 2020 - October 13, 2020
2021-01-23
8 pages
Article/Chapter (Book)
Electronic Resource
English
Temporal and Multi-level Context Attentive GRU Neural Networks
TIBKAT | 2021
|British Library Conference Proceedings | 2001
|Individual Mobility Prediction via Attentive Marked Temporal Point Processes
ArXiv | 2021
|IEEE | 2001
|