Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation

With the vast quantity of GPS data that have been collected from location-based social networks, Point-of-Interest (POI) recommendation aims to predict users’ next locations by learning from their historical check-in trajectories. While Graph Neural Network (GNN)-based models have shown promising re...

Full description

Saved in:
Bibliographic Details
Main Authors: Qiuhan Han, Atsushi Yoshikawa, Masayuki Yamamura
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/9/4979
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850030821335891968
author Qiuhan Han
Atsushi Yoshikawa
Masayuki Yamamura
author_facet Qiuhan Han
Atsushi Yoshikawa
Masayuki Yamamura
author_sort Qiuhan Han
collection DOAJ
description With the vast quantity of GPS data that have been collected from location-based social networks, Point-of-Interest (POI) recommendation aims to predict users’ next locations by learning from their historical check-in trajectories. While Graph Neural Network (GNN)-based models have shown promising results in this field, they typically construct single-layer graphs that fail to capture the hierarchical nature of human mobility patterns. To address this limitation, we propose a novel Hierarchical Graph Learning (HGL) framework that models POI relationships at multiple scales. Specifically, we construct a three-level graph structure: a base-level graph capturing direct POI transitions, a region-level graph modeling area-based interactions through spatio-temporal clustering, and a global-level graph representing category-based patterns. To effectively utilize this hierarchical structure, we design a cross-layer information propagation mechanism that enables bidirectional message passing between different levels, allowing the model to capture both fine-grained POI interactions and coarse-grained mobility patterns. Compared to traditional models, our hierarchical structure improves cold-start robustness and achieves superior performance on real-world datasets. While the incorporation of multi-layer attention and clustering introduces moderate computational overhead, the cost remains acceptable for offline recommendation contexts.
format Article
id doaj-art-a7b6c9a412cc4b0d8f73b49e36c6c967
institution DOAJ
issn 2076-3417
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-a7b6c9a412cc4b0d8f73b49e36c6c9672025-08-20T02:59:07ZengMDPI AGApplied Sciences2076-34172025-04-01159497910.3390/app15094979Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest RecommendationQiuhan Han0Atsushi Yoshikawa1Masayuki Yamamura2Department of Computer Science, School of Computing, Institute of Science Tokyo, Tokyo 152-8550, JapanSchool of Engineering, Kanto Gakuin University, Yokohama 236-0037, JapanDepartment of Computer Science, School of Computing, Institute of Science Tokyo, Tokyo 152-8550, JapanWith the vast quantity of GPS data that have been collected from location-based social networks, Point-of-Interest (POI) recommendation aims to predict users’ next locations by learning from their historical check-in trajectories. While Graph Neural Network (GNN)-based models have shown promising results in this field, they typically construct single-layer graphs that fail to capture the hierarchical nature of human mobility patterns. To address this limitation, we propose a novel Hierarchical Graph Learning (HGL) framework that models POI relationships at multiple scales. Specifically, we construct a three-level graph structure: a base-level graph capturing direct POI transitions, a region-level graph modeling area-based interactions through spatio-temporal clustering, and a global-level graph representing category-based patterns. To effectively utilize this hierarchical structure, we design a cross-layer information propagation mechanism that enables bidirectional message passing between different levels, allowing the model to capture both fine-grained POI interactions and coarse-grained mobility patterns. Compared to traditional models, our hierarchical structure improves cold-start robustness and achieves superior performance on real-world datasets. While the incorporation of multi-layer attention and clustering introduces moderate computational overhead, the cost remains acceptable for offline recommendation contexts.https://www.mdpi.com/2076-3417/15/9/4979next POI recommendationgraph neural networksspatio-temporal data mining
spellingShingle Qiuhan Han
Atsushi Yoshikawa
Masayuki Yamamura
Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation
Applied Sciences
next POI recommendation
graph neural networks
spatio-temporal data mining
title Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation
title_full Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation
title_fullStr Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation
title_full_unstemmed Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation
title_short Hierarchical Graph Learning with Cross-Layer Information Propagation for Next Point of Interest Recommendation
title_sort hierarchical graph learning with cross layer information propagation for next point of interest recommendation
topic next POI recommendation
graph neural networks
spatio-temporal data mining
url https://www.mdpi.com/2076-3417/15/9/4979
work_keys_str_mv AT qiuhanhan hierarchicalgraphlearningwithcrosslayerinformationpropagationfornextpointofinterestrecommendation
AT atsushiyoshikawa hierarchicalgraphlearningwithcrosslayerinformationpropagationfornextpointofinterestrecommendation
AT masayukiyamamura hierarchicalgraphlearningwithcrosslayerinformationpropagationfornextpointofinterestrecommendation