Distance-Dependent competitive dynamics of high-speed rail: A comparative machine learning approach for predicting mode choice and market share redistribution

• The study models the impact of HSR on other modes across various distances. • Road transport retains market dominance in short-distance (0-300 km) segments. • HSR triggers a substantial road-to-rail shift in medium-haul (300–500 km) trips. • HSR replaces air travel as the dominant mode in upper-medium (500-800 km) routes. • SHAP analysis reveals non-linear, distance-dependent choice behaviors. The introduction of high-speed rail (HSR) fundamentally alters intercity transport equilibrium, yet its competitive impact exhibits significant heterogeneity across travel distances and operational configurations. This study establishes a rigorous comparative framework to forecast mode choice behavior and market share redistribution along Thailand’s planned Bangkok-Chiang Mai corridor. Utilizing Stated Preference (SP) data from 3,139 respondents, we benchmark a baseline multinomial logit (MNL) model against five advanced machine learning (ML) algorithms. Empirical results demonstrate that ensemble tree-based models significantly outperform both MNL and deep neural networks, with CatBoost emerging as the superior classifier. Integrating SHapley Additive exPlanations (SHAP) with a comprehensive sensitivity analysis, this research uncovers distinct, distance-dependent competitive regimes and non-linear "threshold effects." In short-distance segments (0–300 km), HSR demand is primarily governed by price sensitivity. In medium-distance segments (300–500 km), a significant "road-to-rail" shift is observed, with demand saturating at an optimal operational velocity of 250 km/h. For longer distances (500–800 km), HSR achieves market dominance as a formidable air transport substitute, provided it breaches a critical 300 km/h tipping point. However, the findings reveal a fragile equilibrium: an "Air Rebound" effect occurs if HSR fares exceed a rigid price ceiling, causing demand to surge back to aviation. These results suggest that a tiered HSR service strategy—balancing high-velocity performance for long hauls with price competitiveness for regional segments—is essential to maximize commercial viability and social sustainability within a multi-modal ecosystem.

Authors

Institutions

Publication Details

Journal
Results in Engineering
Published
2026-03-20
DOI
https://doi.org/10.1016/j.rineng.2026.110196
Primary Topic
Transport and Economic Policies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Distance-Dependent competitive dynamics of high-speed rail: A comparative machine learning approach for predicting mode choice and market share redistribution

A. Srikaew, Sutthipong Meeyai, Panumas Sukwong, Chatchawan Simalorit
Results in Engineering
Transport and Economic Policies
article

Distance-Dependent competitive dynamics of high-speed rail: A comparative machine learning approach for predicting mode choice and market share redistribution

A. Srikaew, Sutthipong Meeyai, Panumas Sukwong, Chatchawan Simalorit
article en

Abstract

• The study models the impact of HSR on other modes across various distances. • Road transport retains market dominance in short-distance (0-300 km) segments. • HSR triggers a substantial road-to-rail shift in medium-haul (300–500 km) trips. • HSR replaces air travel as the dominant mode in upper-medium (500-800 km) routes. • SHAP analysis reveals non-linear, distance-dependent choice behaviors. The introduction of high-speed rail (HSR) fundamentally alters intercity transport equilibrium, yet its competitive impact exhibits significant heterogeneity across travel distances and operational configurations. This study establishes a rigorous comparative framework to forecast mode choice behavior and market share redistribution along Thailand’s planned Bangkok-Chiang Mai corridor. Utilizing Stated Preference (SP) data from 3,139 respondents, we benchmark a baseline multinomial logit (MNL) model against five advanced machine learning (ML) algorithms. Empirical results demonstrate that ensemble tree-based models significantly outperform both MNL and deep neural networks, with CatBoost emerging as the superior classifier. Integrating SHapley Additive exPlanations (SHAP) with a comprehensive sensitivity analysis, this research uncovers distinct, distance-dependent competitive regimes and non-linear "threshold effects." In short-distance segments (0–300 km), HSR demand is primarily governed by price sensitivity. In medium-distance segments (300–500 km), a significant "road-to-rail" shift is observed, with demand saturating at an optimal operational velocity of 250 km/h. For longer distances (500–800 km), HSR achieves market dominance as a formidable air transport substitute, provided it breaches a critical 300 km/h tipping point. However, the findings reveal a fragile equilibrium: an "Air Rebound" effect occurs if HSR fares exceed a rigid price ceiling, causing demand to surge back to aviation. These results suggest that a tiered HSR service strategy—balancing high-velocity performance for long hauls with price competitiveness for regional segments—is essential to maximize commercial viability and social sustainability within a multi-modal ecosystem.

Results in EngineeringVol. 30
Suranaree University of Technology (TH)
Openalex Percentile: Top 13%
Transport and Economic Policies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.