STF-MgNet: A Framework for Numerical Analysis and Result Visualization

STF-MgNet is an integrated, open-source framework designed to streamline the research and application of the STF-MgNet architecture. It provides a comprehensive suite of tools that facilitates the entire workflow, from numerical analysis and model evaluation to the generation and visualization of predicted moment magnitrude by using the early stage of STFs. notice: Because syn_Mw_npy_test\syn_train_STF.npy and scardec_usgs_obs\obs_train_STF.npy are too big for git to upload, please contact us. MgNet_out_selected_events_predicted_Mws.py gives out the predicted magnitudes for the selected 4 events based on different percentages of rupture onsets. MgNet_predicted_acc_rupture_onset_process.py outputs the magnitude-estimation accuracies of MgNet based on different percentages of rupture onsets. MgNet_characteristic_time_determine_Mw.py analyzes the MgNet's characteristic time to determine the final magnitude of large earthquakes at the stage when magnitude-estimation accuracy is stable.