Fig. 7

Simulation results of the proposed model for different picking absolute error bins on the testing data set. a–c indicates the results about RMSE, MAE, and ϭ when time (\({t}_{1}\)) is at 3 s. The black dashed lines indicate the results of the model with P-arrivals picked manually. The gray dashed lines indicate the ratio of increased error levels on the magnitude estimation performance from different picking absolute error bins compared with the results with manual P-arrivals. d–f indicates the error curves of the model for different picking absolute error bins with time (\({t}_{1}\)) increasing. Different colors indicate different picking absolute error bins