VOLUME 19 (Supplement)

SciEnggJ%202026%20Special%20Issue%20265 280 Bautista%20et%20al

SciEnggJ 19 (Supplement) 488-499
available online: 19 November 2026
DOI: https://doi.org/10.54645/202619SupJUC-99

*Corresponding author
Email Address: mmvallejos@bpsu.edu.ph
Date received: 04 April 2026
Dates revised: 14 August 2026;
12 September 2026; 21 September 2026
Date accepted: 21 September 2026

ARTICLE

Segmenting freight operations for route choice analysis: A cluster analysis of truck transport in Bataan, Philippines

Maryrose M. Vallejos*1,2, and Alexis M. Fillone1

1Gokongwei College of Engineering, De La Salle University, Manila 1004, Philippines

2College of Engineering, Bataan Peninsula State University, Balanga City, Bataan 2100, Philippines

KEYWORDS: freight transport; K-means clustering; freight segmentation; truck operations; willingness to pay; toll infrastructure; Bataan–Cavite Interlink Bridge

Freight operations differ in trip conditions, transport costs, shipment characteristics, organizational arrangements, and willingness to pay for new infrastructure. This study applied K-means clustering to the most recent trips reported by 358 freight respondents in Bataan, Philippines, to identify operational segments relevant to the proposed Bataan–Cavite Interlink Bridge (BCIB). Seven standardized variables were analyzed: heavy-goods-vehicle route distance, reported travel time, toll cost, fuel cost, cargo weight, cargo value, and direct stated BCIB willingness to pay (WTP). Candidate solutions with two to eight clusters were evaluated using internal-validation criteria and stability assessments. A three-cluster solution was retained based on validation performance, parsimony, stability, and interpretability. The retained solution had the highest average silhouette coefficient among candidates (0.218). Across 200 repeated 80% subsamples, the mean adjusted Rand index (ARI) was 0.931; 100 random-seed runs produced identical partitions (ARI = 1.000). These findings support a reproducible exploratory segmentation, although separation among freight operations was limited. The identified segments were Short-Haul Lower-Cost Freight Operators (43.0%), High-Value Hazardous Full-Truckload (FTL) Freight Operators (3.6%), and Long-Haul High-Cost Heavy-Freight Operators (53.4%). The specialized high-value segment had the highest cluster-level silhouette coefficient; all 13 observations involved hazardous cargo and full-truckload operations, and the segment had the highest mean direct BCIB WTP. External profiling identified associations with commodity type, hazardous status, truckload type, axle count, truck configuration, truck ownership, and route decision authority. The findings demonstrate interpretable heterogeneity in freight operations and provide an exploratory basis for considering differentiated freight-user profiles in BCIB corridor and toll planning.

© 2026 SciEnggJ
Philippine-American Academy of Science and Engineering