Detecting earthquakes using AI

Abstract

This project employs a quantitative approach to earthquake detection by integrating Artificial Intelligence (AI) and Machine Learning (ML) techniques with seismic data analysis, focusing on the seismic landscape of Malta. Traditional seismology methods face significant challenges in Malta due to the high population density and construction activity, contributing to complex and noisy seismic data. This study aims to address these challenges by developing and evaluating various AI models, including K‐Nearest Neighbours (KNN), Random Forest (RF) and Long Short‐Term Memory (LSTM) networks to accurately detect earthquakes. The research approach involves training these models on seismic data provided by the Seismic Monitoring & Research Group within the Department of Geosciences in the Faculty of Science at the University of Malta (SMRG), encompassing fourteen years of earthquake records. The study systematically evaluates the impact of various data preprocessing techniques, such as noise reduction, filtering and data augmentation, on model performance. Additionally, the research assesses the generalisability of the models across different geographic regions and examines the relationship between dataset size and model performance. The findings indicate that KNN achieved the highest accuracy and F‐score at 99.1%, followed closely by RF with 98.7%, with LSTM showing a lower performance at 92.6%. The study highlights the change in results achieved when utilising various data quality and preprocessing techniques. The models’ generalisability was tested on international data, revealing a significant drop in performance, underscoring the need for region‐specific models or a training dataset which includes data from the seismic stations the model is to be used on. In conclusion, the integration of AI and ML techniques in seismology presents a significant advancement for earthquake detection in Malta. The insights gained from this research contribute to enhancing earthquake monitoring and response systems in Malta, ensuring a safer environment for the Maltese population. This project provides a foundation for future research in AI‐based seismology, emphasising the need for ongoing innovation and collaboration.