Analyzing historical seismic data for region-specific earthquake prediction through deep neural networks

Abstract

This study considers the problem of improving the accuracy of earthquake forecasting in Kazakhstan using deep learning methods. Special attention is paid to forecasting in zones with increased seismic activity, which is a unique feature of the regional plan. The main goal of the work is to develop a model based on the architecture of deep neural networks with direct propagation of signals to analyze historical data containing the time of occurrence, geographic coordinates and magnitude of earthquakes. This model is used as a basis for classification and regression, allowing to evaluate the level of agreement between the predicted values and those obtained. Evidence of earthquake prediction capability is demonstrated using this methodology, achieving 86% accuracy in magnitude classification, and a Mean Squared Error of 0.22 in forecasting the magnitude itself in regression utilizing computational power as well as advanced neural network techniques. The significance of this study is that it contributes to the development of processes and methods that promote the application of deep learning in seismology to improve the accuracy and efficiency of diagnosis and prevention of natural disasters, also opens new perspectives in this direction.

More by Marat Nurtas

• Two-stage synthetic-to-real transfer learning for automated mammography report generation using vision-language modelsScientific Reports · 2026
• Macro-scale earthquake forecasting under class imbalance in Central AsiaJournal Of Big Data · 2026
• Enhanced Fine-Tuning Optimization for Real-Time Fire and Smoke Detection Using a YOLOv9 Architecture with a Comprehensive Ablation AnalysisEngineering Technology & Applied Science Research · 2026 · Citations: 1
• Spatially honest validation of machine-learning prospectivity maps for karst bauxite in KazakhstanResearch Square · 2026