Abstract
The field of aftershock prediction has evolved over the years, transitioning from traditional statistical models to more advanced machine learning and deep learning methodologies. The USGS relies on the Reasenberg and Jones (1989, 1994) statistical model which subsequent researchers have refined to adapt to various tectonic regimes globally. Meanwhile, DeVries et al. (2018) explored the potential of Deep Learning in predicting aftershock locations, albeit facing critique for potentially over-complicating the issue. Other researchers have applied machine learning algorithms, such as in predicting aftershock patterns following the Kermanshah Earthquake in Iran, demonstrating the capability of machine learning to outperform traditional Coulomb maps. The dynamic discourse among these varying methodologies highlights the ongoing efforts to improve aftershock prediction, aiming at better preparedness and response strategies in seismic-prone regions.