Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering

Next point-of-interest (POI) recommendation provides users with location suggestions that they may be interested in, allowing them to explore their surroundings. Existing sequence-based or graph-based POI recommendation methods have matured in capturing spatiotemporal information; however, POI recom...

Full description

Saved in:
Bibliographic Details
Main Authors: Jiubing Chen, Haoyu Wang, Jianxin Shang, Chaomurilige
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Mathematics
Subjects:
Online Access:https://www.mdpi.com/2227-7390/12/22/3592
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846153064996667392
author Jiubing Chen
Haoyu Wang
Jianxin Shang
Chaomurilige
author_facet Jiubing Chen
Haoyu Wang
Jianxin Shang
Chaomurilige
author_sort Jiubing Chen
collection DOAJ
description Next point-of-interest (POI) recommendation provides users with location suggestions that they may be interested in, allowing them to explore their surroundings. Existing sequence-based or graph-based POI recommendation methods have matured in capturing spatiotemporal information; however, POI recommendation methods based on large language models (LLMs) focus more on capturing sequential transition relationships. This raises an unexplored challenge: how to leverage LLMs to better capture geographic contextual information. To address this, we propose interpretable embeddings for next point-of-interest recommendation via large language model question–answering, named QA-POI, which transforms the POI recommendation task into obtaining interpretable embeddings via LLM prompts, followed by lightweight MLP fine-tuning. We introduce question–answer embeddings, which are generated by asking LLMs yes/no questions about the user’s trajectory sequence. By asking spatiotemporal questions about the trajectory sequence, we aim to extract as much spatiotemporal information from the LLM as possible. During training, QA-POI iteratively selects the most valuable subset of potential questions from a set of questions to prompt the LLM for the next POI recommendation. It is then fine-tuned for the next POI recommendation task using a lightweight Multi-Layer Perceptron (MLP). Extensive experiments on two datasets demonstrate the effectiveness of our approach.
format Article
id doaj-art-51dbaa99b07d4e60a78d76d3adf99d9f
institution Kabale University
issn 2227-7390
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Mathematics
spelling doaj-art-51dbaa99b07d4e60a78d76d3adf99d9f2024-11-26T18:11:57ZengMDPI AGMathematics2227-73902024-11-011222359210.3390/math12223592Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–AnsweringJiubing Chen0Haoyu Wang1Jianxin Shang2Chaomurilige3School of Statistics, Jilin University of Finance and Economics, Changchun 130117, ChinaBig Data and Network Management Center, Jilin University, Changchun 130012, ChinaSchool of Information and Technology, Northeast Normal University, Changchun 130024, ChinaKey Laboratory of Ethnic Language Intelligent Analysis and Security Governance, Ministry of Education, Minzu University of China, Haidian District, Beijing 100081, ChinaNext point-of-interest (POI) recommendation provides users with location suggestions that they may be interested in, allowing them to explore their surroundings. Existing sequence-based or graph-based POI recommendation methods have matured in capturing spatiotemporal information; however, POI recommendation methods based on large language models (LLMs) focus more on capturing sequential transition relationships. This raises an unexplored challenge: how to leverage LLMs to better capture geographic contextual information. To address this, we propose interpretable embeddings for next point-of-interest recommendation via large language model question–answering, named QA-POI, which transforms the POI recommendation task into obtaining interpretable embeddings via LLM prompts, followed by lightweight MLP fine-tuning. We introduce question–answer embeddings, which are generated by asking LLMs yes/no questions about the user’s trajectory sequence. By asking spatiotemporal questions about the trajectory sequence, we aim to extract as much spatiotemporal information from the LLM as possible. During training, QA-POI iteratively selects the most valuable subset of potential questions from a set of questions to prompt the LLM for the next POI recommendation. It is then fine-tuned for the next POI recommendation task using a lightweight Multi-Layer Perceptron (MLP). Extensive experiments on two datasets demonstrate the effectiveness of our approach.https://www.mdpi.com/2227-7390/12/22/3592point of interestsequential recommendationlarge language modelsspatiotemporal
spellingShingle Jiubing Chen
Haoyu Wang
Jianxin Shang
Chaomurilige
Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering
Mathematics
point of interest
sequential recommendation
large language models
spatiotemporal
title Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering
title_full Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering
title_fullStr Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering
title_full_unstemmed Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering
title_short Interpretable Embeddings for Next Point-of-Interest Recommendation via Large Language Model Question–Answering
title_sort interpretable embeddings for next point of interest recommendation via large language model question answering
topic point of interest
sequential recommendation
large language models
spatiotemporal
url https://www.mdpi.com/2227-7390/12/22/3592
work_keys_str_mv AT jiubingchen interpretableembeddingsfornextpointofinterestrecommendationvialargelanguagemodelquestionanswering
AT haoyuwang interpretableembeddingsfornextpointofinterestrecommendationvialargelanguagemodelquestionanswering
AT jianxinshang interpretableembeddingsfornextpointofinterestrecommendationvialargelanguagemodelquestionanswering
AT chaomurilige interpretableembeddingsfornextpointofinterestrecommendationvialargelanguagemodelquestionanswering