Adapting large language models for urban trip purpose inference
A domain-informed textual representation and fine-tuning framework
Bibliographic Data
| ID | 24227228 |
|---|---|
| Authors | Chengzhi Zhang (0000-0002-7367-0334), Shaoying Li (0000-0002-4703-5660), Xiaoyi Ma (0000-0002-3631-8673), Xianlong Chen, Zhou Huang (0000-0002-1255-1913), Zixuan Bai, Zheng Gong (0009-0007-7711-2376) |
| Year | 2026 |
| Volume | 179 |
| Pages | 107680 |
| Publication date | 2026-12-01 |
| Open Access | No |
| Type | ARTICLE |
| Venue | Cities (JOURNAL) |
| Journal identifiers | ISSN: 0264-2751 • E-ISSN: 1873-6084 |
| Publisher | Elsevier BV (PUBLISHER) |
| DOI | 10.1016/j.cities.2026.107680 |
| OpenAlex | W7221803238 |
| Language | EN |
| References cited | 58 |
Understanding why people travel is fundamental to activity-based travel demand modelling and transport policy analysis. However, conventional trip purpose inference methods typically represent mobility contexts using predefined numerical or categorical features, which may discard semantic information contained in raw contextual data, thereby limiting their ability to represent and distinguish diverse travel-related contexts. To address this gap, this study proposes TripPurpose LLM, a domain-informed trip purpose inference framework based on large language models (LLMs). The framework integrates traveller attributes, departure time, transport mode, and semantic information from nearby points of interest (POIs) into a unified textual representation, incorporates transport domain knowledge through structured chain-of-thought (CoT) guidance, and adapts pretrained LLMs using low-rank adaptation (LoRA) fine-tuning. Experiments used 10,513 survey-based trip records from Guangzhou, China. Within the TripPurpose LLM framework, retaining the original POI names increased the accuracy of Qwen3-8B from 0.695 to 0.745 and its Macro-F1 from 0.610 to 0.681, with both metrics exceeding those of structured-feature machine-learning baselines and classifiers using term frequency–inverse document frequency (TF-IDF) features or fixed MacBERT embeddings. This finding highlights the value of contextual semantic modelling for trip purpose inference. Across ten LLM backbones, TripPurpose LLM combining LoRA fine-tuning with CoT guidance consistently achieved the strongest performance among the evaluated settings, with accuracy ranging from 0.71 to 0.76. This pattern demonstrates the framework's effectiveness across different model families and parameter scales. This study provides an integrative framework for inferring trip purposes from travel records, thereby supporting travel-demand analysis, transport planning, and policy evaluation.
activity-based travel demand modeling · chain-of-thought prompting · departure time · large language models · low-rank adaptation · points of interest · transport modes · Transportation planning · travel survey data · Human Mobility and Location-Based Analysis · Traffic Prediction and Management Techniques · Transportation Planning and Optimization
Understanding individual human mobility patterns
Applying mobile phone data to travel behaviour research
Dynamic population mapping using mobile phone data
Origin–destination trips by purpose and time of day inferred from mobile phone data
A systematic analysis of performance measures for classification tasks
Growth, innovation, scaling, and the pace of life in cities
Understanding spatiotemporal trip purposes of urban micro-mobility from the lens of dockless e-scooter sharing
Exploring large language models for human mobility prediction under public events
Research note
Portraying business district vibrancy with mobile phone data and optimal parameters-based geographical detector model
How urban land use influences commuting flows in Wuhan, Central China
Leveraging LLM-based multi-agent simulations to boost participatory design education
Detecting home location and trip purposes for cardholders by mining smart card transaction data in Beijing subway
The life cycle concept as a tool for travel research
Advancements in the application of large language models in urban studies
What About People in Regional Science
Big data for big issues
An approach to imputing destination activities for inclusion in measures of bicycle accessibility
Uncovering the spatial patterns and internal vitality heterogeneity of bedroom communities
The role of planning policy in the evolution of the spatial structure of the Guangzhou metropolitan area in China
The six dimensions of built environment on urban vitality
GeoAvatar
Urban planning in the age of large language models
A wildfire exposure risk assessment framework combining LLM and spatial-temporal knowledge hypergraphs
Multifaceted associations between built environments and POI visit patterns by trip purposes
Making a city
The Use of the Life-Style Concept in Travel Demand Models
Mobility Research in the Age of the Smartphone
| Highly cited | No |
|---|