Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Forecasting Unseen Points of Interest Visits Using Context and Proximity Priors

Ziyao Li, Shang-Ling Hsu, Cyrus Shahabi

TL;DR

This work proposes a model designed to predict a new POI outside the training data as long as its context is aligned with the user’s interests, and demonstrates that this model outperforms baseline methods that do not account for semantic contexts.

Abstract

Understanding human mobility behavior is crucial for numerous applications, including crowd management, location-based recommendations, and the estimation of pandemic spread. Machine learning models can predict the Points of Interest (POIs) that individuals are likely to visit in the future by analyzing their historical visit patterns. Previous studies address this problem by learning a POI classifier, where each class corresponds to a POI. However, this limits their applicability to predict a new POI that was not in the training data, such as the opening of new restaurants. To address this challenge, we propose a model designed to predict a new POI outside the training data as long as its context is aligned with the user's interests. Unlike existing approaches that directly predict specific POIs, our model first forecasts the semantic context of potential future POIs, then combines this with a proximity-based prior probability distribution to determine the exact POI. Experimental results on real-world visit data demonstrate that our model outperforms baseline methods that do not account for semantic contexts, achieving a 17% improvement in accuracy. Notably, as new POIs are introduced over time, our model remains robust, exhibiting a lower decline rate in prediction accuracy compared to existing methods.

Forecasting Unseen Points of Interest Visits Using Context and Proximity Priors

TL;DR

This work proposes a model designed to predict a new POI outside the training data as long as its context is aligned with the user’s interests, and demonstrates that this model outperforms baseline methods that do not account for semantic contexts.

Abstract

Understanding human mobility behavior is crucial for numerous applications, including crowd management, location-based recommendations, and the estimation of pandemic spread. Machine learning models can predict the Points of Interest (POIs) that individuals are likely to visit in the future by analyzing their historical visit patterns. Previous studies address this problem by learning a POI classifier, where each class corresponds to a POI. However, this limits their applicability to predict a new POI that was not in the training data, such as the opening of new restaurants. To address this challenge, we propose a model designed to predict a new POI outside the training data as long as its context is aligned with the user's interests. Unlike existing approaches that directly predict specific POIs, our model first forecasts the semantic context of potential future POIs, then combines this with a proximity-based prior probability distribution to determine the exact POI. Experimental results on real-world visit data demonstrate that our model outperforms baseline methods that do not account for semantic contexts, achieving a 17% improvement in accuracy. Notably, as new POIs are introduced over time, our model remains robust, exhibiting a lower decline rate in prediction accuracy compared to existing methods.

Paper Structure

This paper contains 17 sections, 13 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Trajectory Synthesis
  4. Next Visit Forecasting
  5. Problem Formulation
  6. Background
  7. Method
  8. Joint Probability Approximation
  9. Semantic Context Prediction
  10. Visit Sequence Encoder
  11. Semantic Context Classification
  12. Proximity Prior Computation
  13. Experiments
  14. Experimental Setup
  15. Results and Discussion
  16. ...and 2 more sections

Figures (4)

  • Figure 1: Our proposed model.
  • Figure 2: The proximity priors histogram representing the number of instances where the distance between consecutive user locations falls within specific distance intervals (in kilometers).
  • Figure 3: The training, testing, validation, and unseen POI evaluation split.
  • Figure 4: The comparison of MobTCast and our model accuracy as the percentage of unseen POIs increases.