Evaluation of Earthquake Hazard Risk Level Based on Random Forest

Abstract

Earthquakes, highly destructive natural disasters, pose significant challenges to China, a country situated at the intersection of two major seismic belts. The risk of earthquake disasters has been escalating due to population density and rapid economic development, threatening the safety of people's lives and the sustainability of the national economy. Therefore, rapid and accurate earthquake risk assessment is crucial for disaster prevention and mitigation. With the advancement of artificial intelligence technology, machine learning algorithms have become vital tools in earthquake science research. This study aims to develop a nationwide earthquake risk assessment model, utilizing earthquake data from 2005 to 2020. Through preprocessing techniques such as normalization, discretization, and type conversion, combined with Spearman correlation analysis to select key indicators. After training and testing with BP neural network, SVM, and Random Forest models, the Random Forest model demonstrated the best performance in key metrics such as accuracy, precision, recall, and F1 Score, proving its superior classification capability. After a comprehensive evaluation, the Random Forest model was chosen as the preferred model for earthquake risk assessment, ensuring the accuracy and reliability of the assessment.