Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Where to Move Next: Zero-shot Generalization of LLMs for Next POI Recommendation

Shanshan Feng, Haoming Lyu, Caishun Chen, Yew-Soon Ong

TL;DR

This work tackles zero-shot next-POI recommendation by leveraging large language models through a novel prompting framework, LLMmove, which integrates long-term and recent user preferences, geospatial distance, and sequential transitions. Experiments on NYC and TKY show that LLMmove achieves state-of-the-art zero-shot performance and outperforms geography-agnostic baselines, underscoring the value of structured prompts for spatial-temporal reasoning. The study also uncovers limitations in geographic context understanding and sensitivity to candidate ordering, highlighting the need for robust spatial reasoning in LLM-based mobility systems. Overall, the paper contributes a first zero-shot POI framework with open-source resources and provides insights into how to harness LLMs for city-scale mobility recommendations.

Abstract

Next Point-of-interest (POI) recommendation provides valuable suggestions for users to explore their surrounding environment. Existing studies rely on building recommendation models from large-scale users' check-in data, which is task-specific and needs extensive computational resources. Recently, the pretrained large language models (LLMs) have achieved significant advancements in various NLP tasks and have also been investigated for recommendation scenarios. However, the generalization abilities of LLMs still are unexplored to address the next POI recommendations, where users' geographical movement patterns should be extracted. Although there are studies that leverage LLMs for next-item recommendations, they fail to consider the geographical influence and sequential transitions. Hence, they cannot effectively solve the next POI recommendation task. To this end, we design novel prompting strategies and conduct empirical studies to assess the capability of LLMs, e.g., ChatGPT, for predicting a user's next check-in. Specifically, we consider several essential factors in human movement behaviors, including user geographical preference, spatial distance, and sequential transitions, and formulate the recommendation task as a ranking problem. Through extensive experiments on two widely used real-world datasets, we derive several key findings. Empirical evaluations demonstrate that LLMs have promising zero-shot recommendation abilities and can provide accurate and reasonable predictions. We also reveal that LLMs cannot accurately comprehend geographical context information and are sensitive to the order of presentation of candidate POIs, which shows the limitations of LLMs and necessitates further research on robust human mobility reasoning mechanisms.

Where to Move Next: Zero-shot Generalization of LLMs for Next POI Recommendation

TL;DR

This work tackles zero-shot next-POI recommendation by leveraging large language models through a novel prompting framework, LLMmove, which integrates long-term and recent user preferences, geospatial distance, and sequential transitions. Experiments on NYC and TKY show that LLMmove achieves state-of-the-art zero-shot performance and outperforms geography-agnostic baselines, underscoring the value of structured prompts for spatial-temporal reasoning. The study also uncovers limitations in geographic context understanding and sensitivity to candidate ordering, highlighting the need for robust spatial reasoning in LLM-based mobility systems. Overall, the paper contributes a first zero-shot POI framework with open-source resources and provides insights into how to harness LLMs for city-scale mobility recommendations.

Abstract

Next Point-of-interest (POI) recommendation provides valuable suggestions for users to explore their surrounding environment. Existing studies rely on building recommendation models from large-scale users' check-in data, which is task-specific and needs extensive computational resources. Recently, the pretrained large language models (LLMs) have achieved significant advancements in various NLP tasks and have also been investigated for recommendation scenarios. However, the generalization abilities of LLMs still are unexplored to address the next POI recommendations, where users' geographical movement patterns should be extracted. Although there are studies that leverage LLMs for next-item recommendations, they fail to consider the geographical influence and sequential transitions. Hence, they cannot effectively solve the next POI recommendation task. To this end, we design novel prompting strategies and conduct empirical studies to assess the capability of LLMs, e.g., ChatGPT, for predicting a user's next check-in. Specifically, we consider several essential factors in human movement behaviors, including user geographical preference, spatial distance, and sequential transitions, and formulate the recommendation task as a ranking problem. Through extensive experiments on two widely used real-world datasets, we derive several key findings. Empirical evaluations demonstrate that LLMs have promising zero-shot recommendation abilities and can provide accurate and reasonable predictions. We also reveal that LLMs cannot accurately comprehend geographical context information and are sensitive to the order of presentation of candidate POIs, which shows the limitations of LLMs and necessitates further research on robust human mobility reasoning mechanisms.
Paper Structure (19 sections, 1 equation, 2 figures, 5 tables)

This paper contains 19 sections, 1 equation, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Next-POI Recommendation
  4. LLM-Based Recommender Systems
  5. LLMs for User Mobility Patterns
  6. Problem Statement
  7. Methodology
  8. Construction of Input Data
  9. Statement of Requirements
  10. Recommendation and Explanation
  11. Experiments
  12. Experimental Setup
  13. Datasets
  14. Evaluated Methods
  15. Evaluation Metrics
  16. ...and 4 more sections

Figures (2)

  • Figure 1: An example of the next POI recommendation task. Each POI is associated with geographical coordinates and category information. Given the recent trajectory, e.g., $\{l_1, l_2, l_3, l_4\}$, it aims to predict POIs to visit subsequently.
  • Figure 2: The workflow of the designed LLMmove framework and the corresponding prompts.