Technology Category
Machine Learning: Large Multimodal Models (LMMs)Humans and AI: Human-Aware Planning and Behavior PredictionPlanning, Routing, and Scheduling: Planning with Language Models
Can Large Language Models Trigger a Paradigm Shift in Travel Behavior Modeling? Experiences with Modeling Travel Satisfaction
🤖 AI Summary
Large language models (LLMs) exhibit systematic behavioral biases when predicting travel satisfaction, diverging from empirically observed human mobility patterns. Method: This work first systematically identifies and quantifies such behavioral biases in LLM-based travel behavior modeling, and proposes a few-shot bias calibration framework that requires only a small number of labeled samples—without relying on large-scale datasets or strong statistical assumptions. Leveraging zero-shot and few-shot prompting, the method is fine-tuned and evaluated using Shanghai household travel survey data. Results: Under few-shot settings, the calibrated LLM significantly outperforms conventional statistical and machine learning baselines in MSE and MAPE. The study demonstrates that behavior-aware calibration endows LLMs with high-accuracy, strongly generalizable travel satisfaction modeling capabilities even with minimal supervision, establishing a novel paradigm for behavior-driven AI in transportation modeling.
Technology Category
Application Category
📝 Abstract
As a specific domain of subjective well-being, travel satisfaction has attracted much research attention recently. Previous studies primarily use statistical models and, more recently, machine learning models to explore the determinants of travel satisfaction. Both approaches require data from sufficient sample sizes and correct prior statistical assumptions. The emergence of Large Language Models (LLMs) offers a new modeling approach that can overcome the shortcomings of the existing methods. Pre-trained on extensive datasets, LLMs have strong capabilities in contextual understanding and generalization, significantly reducing their dependence on large quantities of task-specific data and stringent statistical assumptions. The primary challenge in applying LLMs lies in addressing the behavioral misalignment between LLMs and human behavior. Using data on travel satisfaction from a household survey in shanghai, this study identifies the existence and source of misalignment and develop methods to address the misalignment issue. We find that the zero-shot LLM exhibits behavioral misalignment, resulting in relatively low prediction accuracy. However, few-shot learning, even with a limited number of samples, allows the model to outperform baseline models in MSE and MAPE metrics. This misalignment can be attributed to the gap between the general knowledge embedded in LLMs and the specific, unique characteristics of the dataset. On these bases, we propose an LLM-based modeling approach that can be applied to model travel behavior using samples of small sizes. This study highlights the potential of LLMs for modeling not only travel satisfaction but also broader aspects of travel behavior.
Problem
Research questions and friction points this paper is trying to address.
Overcoming data size and assumption limits in travel satisfaction modeling
Addressing behavioral misalignment between LLMs and human travel behavior
Developing LLM-based models for small-sample travel behavior analysis
Innovation
Methods, ideas, or system contributions that make the work stand out.
LLMs reduce need for large task-specific datasets
Few-shot learning improves LLM prediction accuracy
LLMs address behavioral misalignment with humans
🔎 Similar Papers
No similar papers found.
