Related Experiment Video
Updated: Jan 12, 2026
16:14
Trajectory Data Analyses for Pedestrian Space-time Activity Study
Published on: February 25, 2013
14.1K
Probabilistic Bayesian learning with long-tail awareness for trajectory-user linking
Haolun Ding1, Zhengwen Fu2, Rong Zhang2
1Engineering Research Center of Intelligent Finance, Ministry of Education, Southwestern University of Finance and Economics, Chengdu, China.
Summary
This study introduces LongTUL, a new method for linking user trajectories in location-based social networks. It effectively addresses the long tail phenomenon, improving accuracy for less active users.
Area of Science:
- GeoAI
- Data Science
- Machine Learning
Background:
- Location-based social networks (LBSN) generate vast user check-in data, enabling mobility pattern analysis.
- Trajectory-User Linking (TUL) aims to associate unlabeled trajectories with their creators, a key task in GeoAI.
- The 'long tail phenomenon' in user check-ins, where some users have many check-ins and others very few, poses a significant challenge for TUL accuracy.
Purpose of the Study:
- To propose a novel probabilistic Bayesian learning solution, LongTUL, to address the long tail issue in Trajectory-User Linking (TUL).
- To improve the accurate association of unlabeled check-in trajectories with their corresponding users, particularly for less active (tail) users.
Main Methods:
- Developed a Check-in Engagement Compromise (CEC) mechanism to balance user participation levels before training.
- Implemented a Probabilistic Trajectory Learning (PTL) procedure using variational Bayes to encode trajectories into a latent space.
- Applied Laplacian approximation to latent representations to mitigate amortization errors and long-tail effects.
- Designed a reweighted classifier for equitable inference between frequent (head) and infrequent (tail) users.
Main Results:
- The proposed LongTUL method demonstrated superior performance compared to existing TUL solutions.
- LongTUL effectively addressed the long tail issue, significantly improving the classification accuracy for tail users.
- Experiments on three real-world datasets validated the effectiveness of the LongTUL approach.
Conclusions:
- LongTUL offers a robust solution for Trajectory-User Linking, particularly in datasets exhibiting the long tail phenomenon.
- The method enhances the understanding of mobility patterns by accurately linking trajectories from all user groups.
- This work contributes to advancing GeoAI by providing a more equitable and accurate approach to user trajectory analysis.
Related Concept Videos
01:12Observational Learning
804
Albert Bandura's observational learning, also known as imitation or modeling, occurs when a person observes and imitates another's behavior. It is a quicker process than operant conditioning. A well-known example is the Bobo doll study, where children who saw an adult acting aggressively towards the doll were more likely to act aggressively when left alone, compared to those who observed a nonaggressive adult. Many psychologists view observational learning as a form of latent learning...
804
01:27Associative Learning
1.2K
Associative learning is a fundamental concept in behavioral psychology, wherein a connection is established between two stimuli or events, leading to a learned response. This process is critical in understanding how behaviors are acquired and modified. Conditioning, the mechanism through which associations are formed, can be divided into two main types: classical conditioning and operant conditioning, each elucidating different aspects of associative learning.
Classical conditioning, also known...
Classical conditioning, also known...
1.2K