Title: Optimal location prediction for emergency stations using machine learning

Authors: Prasham Sheth; Praxal Patel; Priyank Thakkar

Addresses: Computer Science and Engineering Department, Institute of Technology, Nirma University, Ahmedabad, Gujarat, India ' Computer Science and Engineering Department, Institute of Technology, Nirma University, Ahmedabad, Gujarat, India ' Computer Science and Engineering Department, Institute of Technology, Nirma University, Ahmedabad, Gujarat, India

Abstract: Time is a critical aspect in emergency circumstances like medical crises, natural disasters, breaking out of a fire, etc. The average response time of emergency services is on the rise in recent times owing to the growing traffic. This has raised some serious concerns for people's safety. It is easy to perceive that optimally located emergency stations (e.g., ambulance, fire station) can help in these situations by minimising travel time to reach the location of casualty. With this motivation, we propose an approach which employs K-medoids driven by extreme gradient boosting (XGBoost) for predicting optimal locations of emergency stations. The proposed approach is validated on real datasets, namely: New York City, USA 100-metre Grid Coordinates, NYC Taxi Trip Duration, KNYC Metars 2016 and FDNY Firehouse Listing dataset and the results demonstrate that the proposed method reduces normal average response time and allows serving more locations.

Keywords: emergency station; optimal location prediction; OLP; machine learning; XGBoost; K-medoids; average response time.

DOI: 10.1504/IJOR.2025.144319

International Journal of Operational Research, 2025 Vol.52 No.2, pp.230 - 251

Received: 17 Dec 2021
Accepted: 18 Jul 2022
Published online: 07 Feb 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article