The present study aimed to evaluate the performance of six different regression models for earthquake magnitude prediction using seismic data, a novel approach in the field of earthquake prediction. The selected models were K-Nearest Neighbors (KNN), linear regression, decision trees, random forest, support vector regression, neural networks, and gradient boosting. The evaluation of the models was based on various metrics such as mean squared error (MSE), mean absolute error (MAE), R-squared score, and explained variance score. The data was split into training and testing sets to ensure that the models were being evaluated on unseen data and to avoid overfitting. The results showed that the random forest and gradient boosting models performed the best in predicting earthquake magnitudes using seismic data, which is a significant finding in the field of earthquake prediction. These models had the highest R-squared scores, which indicates that they captured a significant portion of the variance in the target variable, magnitude. On the other hand, the neural networks and support vector regression models performed poorly, with negative R-squared scores, suggesting that they were not a good fit for the data. This study provides a novel contribution to the field of earthquake prediction, as it provides valuable insights into the effectiveness of machine learning algorithms for earthquake magnitude prediction using seismic data. The results clearly demonstrate that the random forest and gradient boosting models are the most effective models for this task, which has practical implications for earthquake hazard assessment and disaster management. The results of this study contribute to the growing body of knowledge in the field of earthquake prediction and have the potential to inform future research in this area.

Earthquake Early Warning Systems (EEWS) represent one of the most effective technological solutions for mitigating the impacts of strong ground motion in seismically active regions. This study presents the design, implementation, and comprehensive evaluation of a real-time earthquake early warning system for Izmir-a region in Western Anatolia characterized by complex tectonic structures and high seismic hazard-using multi-station seismic acceleration data. The proposed framework integrates multi-threaded data acquisition, signal preprocessing, Min-Max normalization, and Euclidean distance-based similarity analysis to enable rapid detection of anomalous seismic patterns during the early P-wave phase. The system architecture consists of distributed sensor inputs, centralized real-time processing, similarity-based anomaly detection, and user-oriented visualization and alerting modules. The performance of the system was evaluated using both real and synthetic seismic datasets. Instrumental earthquake catalog from the 12 June 2017 Karaburun (Mw 6.2) and 30 October 2020 Samos (Mw 6.6) earthquakes demonstrate that the system can generate early warnings 18 s and 13 s prior to strong ground shaking, respectively. In addition, synthetic seismic scenarios were employed to assess system robustness under varying noise levels, station configurations, and signal conditions. The results indicate that the proposed framework maintains stable detection performance and low false-positive rates across diverse operational scenarios. The methodology emphasizes computational efficiency and inter-station waveform coherence analysis, providing a lightweight alternative to conventional magnitude-based approaches. By avoiding computationally intensive source inversion, the system achieves low-latency performance while preserving detection reliability. The proposed EEWS demonstrates strong generalization capability, scalability, and practical applicability for real-time deployment in earthquake-prone urban environments.

Abstract The two frequency‐based magnitude proxies currently employed by earthquake early warning systems in California are the predominant and the characteristic periods. These proxies, obtained using simple expressions that are valid for noise‐free monochromatic signals, yield erroneous result. The log‐average period, τ log , introduced in this study, is calculated directly from the actual velocity spectrum of the first few seconds of the seismic record. Using data from South California and Japan consisting of 440 earthquakes whose magnitudes range between 3 and 7.3, it is demonstrated that τ log is better correlated with the catalog magnitude than the predominant period and provides better magnitude assessment than the characteristic period for small magnitudes ( M <4). The average prediction error is reduced with increasing the input interval up to 6 s. It appears that a single linear scaling describes the relation between log( τ log ) and the catalog magnitude for the entire magnitude range studied here.

caused by an earthquake at a given location. Magnitudes are usually determined from measurements of an earthquake's seismic waves as recorded on a seismogram Seismic magnitude scales are used to describe the overall strength or "size" of an earthquake. These are distinguished from seismic intensity scales that categorize the intensity or severity of ground shaking (quaking) caused by an earthquake at a given location. Magnitudes are usually determined from measurements of an earthquake's seismic waves as recorded on a seismogram. Magnitude scales vary K…