Conference Paper
A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction
Abstract
This paper enhances traditional gravity models for trip demand prediction by integrating geographical, economic, social, and travel datasets with machine learning techniques. The study evaluates county-level data from Tennessee and New York and reports large improvements over traditional gravity modeling across explanatory power, absolute error, and commuter-pattern reliability metrics.
Plain-Language Summary
The paper upgrades the classic gravity model used in transportation planning by adding machine learning and richer datasets, making origin-destination trip predictions more accurate.
Why This Paper Matters
The work provides transportation planners with a more reliable forecasting approach while preserving the practical structure of gravity-based demand modeling.
Research Summary
This paper revisits the gravity model, a classic tool for estimating trips between regions. Gravity models are simple and interpretable, but they can miss complex interactions between geography, socioeconomic conditions, and real travel behavior.
The study enhances gravity modeling with machine learning and richer datasets from Tennessee and New York. By combining traditional modeling structure with data-driven learning, the paper aims to improve prediction accuracy without abandoning a familiar transportation-planning framework.
The result is a more flexible demand prediction approach that can support planning decisions for infrastructure development, mobility services, and regional transportation analysis.
Key Contributions
- Integrates geographic, economic, social, and travel data into gravity-model enhancement.
- Uses machine learning to capture nonlinear trip-demand relationships.
- Validates the approach on Tennessee and New York county-level travel data.
- Reports improved performance across R-squared, MAE, and CPC metrics.
Publication Details
- Type
- Conference Paper
- Venue
- IEEE CAI 2025
- Year
- 2025
Authors
Acharya K., Lad M., Sun L., Song H.
Research Topics
Links and Access
Citation
@inproceedings{acharya2025-a-data-driven-approach-to-enhancing-gravity-models-for-trip-demand-prediction,
title = {A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction},
author = {Acharya K. and Lad M. and Sun L. and Song H.},
booktitle = {IEEE CAI 2025},
year = {2025},
doi = {10.1109/CAI64502.2025.00145}
}