A likelihood-free framework for seismic forecasting models

Likelihood-free inference pipeline for estimating parameters of an incomplete ETAS-type seismicity model with short-term aftershock incompleteness (ETASI). The workflow compares observed and simulated earthquake catalogs through normalized space-time-magnitude summary statistics rather than evaluating the ETAS/ETASI likelihood directly. The code is written in Fortran and is organized into two main stages: Preprocessing of the observed catalog. Simulation-based inference using ETASI catalog simulations and summary-statistic matching.