Kamal Acharya

Postdoctoral Research Associate at Baylor University

Back to Publications

Conference Paper

A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction

By Kamal Acharya Mehul Lad Liang Sun Houbing Herbert Song

An IEEE CAI conference paper that enhances classic gravity models for trip-demand prediction with richer regional datasets and machine learning.

2025 IEEE CAI 2025 DOI: 10.1109/CAI64502.2025.00145

Gravity Models Trip Demand Prediction Data-Driven Modeling

DOI Cite

Abstract

Accurate prediction of trips between zones is critical for transportation planning, as it supports resource allocation and infrastructure development across various modes of transport. Although the gravity model has been widely used due to its simplicity, it often inadequately represents the complex factors influencing modern travel behavior. This study introduces a data-driven approach to enhance the gravity model by integrating geographical, economic, social, and travel data from the counties in Tennessee and New York state. Using machine learning techniques, we extend the capabilities of the traditional model to handle more complex interactions between variables. Our experiments demonstrate that machine learning-enhanced models significantly outperform the traditional model. Our results show a 51.48% improvement in R2, indicating a substantial enhancement in the model’s explanatory power. Also, a 63.59% reduction in Mean Absolute Error (MAE) reflects a significant increase in prediction accuracy. Furthermore, a 44.32% increase in Common Part of Commuters (CPC) demonstrates improved prediction reliability. These findings highlight the substantial benefits of integrating diverse datasets and advanced algorithms into transportation models. They provide urban planners and policymakers with more reliable forecasting and decision-making tools.

In brief

What This Work Does

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.

Research impact

Why It Matters

The work provides transportation planners with a more reliable forecasting approach while preserving the practical structure of gravity-based demand modeling.

Paper at a glance

Research Scope

4Research stages
4Methods or application areas
5Future research directions

Enhanced Gravity Modeling Pipeline

1

Collect Regional Data

Combines geographic, economic, social, and travel datasets for county-level demand modeling.

2

Represent OD Pairs

Builds origin-destination features that preserve the practical structure of gravity-based planning.

3

Learn Nonlinear Effects

Uses machine learning to capture relationships that traditional gravity models can miss.

4

Validate Forecasts

Evaluates demand estimates with R-squared, MAE, and CPC across Tennessee and New York.

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 large improvements across R-squared, MAE, and CPC compared with the traditional gravity model.

Method Components

Gravity Model Baseline

Provides an interpretable transportation-planning foundation for origin-destination trip prediction.

Machine Learning Enhancement

Improves prediction by learning nonlinear relationships among geography, economy, society, and travel behavior.

Multi-Region Validation

Tests the approach on Tennessee and New York data to examine regional applicability.

Feature Importance

Analyzes top-ranked predictors such as distance, travel time, population, and economic factors.

Research Gaps

  1. Cross-region generalization
  2. Dynamic travel behavior
  3. Feature availability
  4. Interpretability
  5. Policy sensitivity

Publication Details

Type
Conference Paper
Venue
IEEE CAI 2025
Year
2025
Pages
815-820

Research Topics

Gravity Models Trip Demand Prediction Data-Driven Modeling Machine Learning

Citation

@inproceedings{acharya2025data,
  author={Acharya, Kamal and Lad, Meet and Sun, Liang and Song, Houbing},
  title={A Data-Driven Approach to Enhancing Gravity Models for Trip Demand Prediction},
  booktitle={IEEE CAI 2025},
  year={2025},
  pages={815--820},
  doi={10.1109/CAI64502.2025.00145}
}