Modeling Seismic Signals of an Earthquake through Symbolic Dynamical System: The Case of Surigao Region, Philippines

Abstract

Earthquakes pose threats globally, most especially in developing hotspots such as the Philippines, given that it falls under the Pacific Ring of Fire. One of the most seismically active regions within the country is the Surigao region located in northeast Mindanao, where two large earthquakes hit in December 2023 with 7.4 and 6.8 magnitudes. The outcome of this study underscores the need for accurate earthquake forecasting models in such high-risk regions. Conventional methods for seismic signal modeling usually lose the complexity and nonlinear of seismic datasets. This paper uses symbolic regression with genetic programming as a means to come up with a more accurate and explainable model in estimating earthquake magnitudes in the Surigao area. A technique in machine learning, symbolic regression generates mathematical expressions which best fit empirical data, and it doesn't require a predefined model configuration. The developed model identifies critical parameters relevant to seismic phenomena and accurately estimates the magnitudes of earthquakes within the considered region. The model, such as the magnitude of 3.0, as happening in September 2024. The results obtained in this study contribute to the accuracy of forecasting for earthquakes by providing SR as an efficient tool for studying seismic records. These findings have the potential to further improve this effort in disaster preparedness and risk reduction that saves lives and reduces economic losses in earthquake-prone areas. Collaboration is encouraged with a continued comprehensive revision toward research for further exploration of broader applications in seismic hazard analysis.