A model of machine learning in megathrust earthquake preparedness recommendation system: Case study of Padang City

¹ Information System, Faculty of Pharmacy Science and Technology, Universitas Dharma Andalas, Padang, Indonesia
² Informatics Engineering Department, Institut Bisnis dan Teknologi Pelita Indonesia, Pekanbaru, Indonesia
³ Software Engineering, Faculty of Technology and Business, Institut Teknologi dan Bisnis Diniyyah Lampung, Indonesia
⁴ Information Management, Faculty of Computer Science, Institut Teknologi Bisnis dan Bahasa Dian Cipta Cendikia, Lampung, Indonesia
⁵ Civil Engineering Department, Universitas Andalas, Padang, Indonesia
⁶ Faculty of Computing and Informatics, Universiti Malaysia Sabah, Sabah, Malaysia

^* Corresponding author: ruddykurniawan@eng.unand.ac.id

Abstract

This document offers an in-depth analysis of the use of machine learning (ML) methods in developing a decision-support system for earthquake preparedness, focusing on the high-risk megathrust scenario originating from the Mentawai segment that threatens Padang City, Indonesia. Utilizing a simulated dataset representing geospatial, demographic, and historical seismic indicators, this work implemented a classification model using Random Forest algorithms to identify and map potential risk zones. The methodology encompasses synthetic data generation, rule-based labeling, model training, evaluation, and predictive inference. Results showed a high classification accuracy under idealized data conditions, and a proof-of-concept interface was developed for scenario-based user input. Thse study concludes with recommendations for integrating real-time data to support disaster risk reduction policies.