UCLA

Understanding that large earthquakes can be violent to human beings, a wide variety of seismicity forecasting models are proposed. To perform useful comparisons among these point process models, graphical residual methods are proposed because they visualize and highlight when and where a given model does not agree closely with either the observed seismicity or another model. In this paper, we work with the one-day observational and forecasting earthquake data from Southern California in the entire year of 2017. Our goal is to utilized Voronoi analysis and Super-thinned analysis to evaluate and compare the Epidemic- Type Aftershock Sequence model (ETAS) and the Modified Nonparametric Hawkes Point Process model (Modified MISD). Essentially, our results suggest that the Modified MISD Model returns comparatively better predictions and better goodness of fit.