Fig. 3: RMSE distribution comparison of log(RotD50) for ASK-14 GMM, RF, and DNN estimations on the synthetic result. | Communications Earth & Environment

Fig. 3: RMSE distribution comparison of log(RotD50) for ASK-14 GMM, RF, and DNN estimations on the synthetic result.

From: A machine learning estimator trained on synthetic data for real-time earthquake ground-shaking predictions in Southern California

Fig. 3

RMSE distribution of log(RotD50) for the ASK-14 GMM (green boxplots), RF (blue boxplots), and DNN (orange boxplots) estimations against the synthetic results shown per magnitude bin for a T = 2s, b T = 3s, c T = 5s, d T = 10s. The median value of the RMSE, marked with a circle for each boxplot, is reduced up to 45% for the ML estimates relative to ASK-14. The mean of the median RMSE value for all magnitudes is summarized in Supplementary Table 2. Boxplot bars represent the first and third quartiles of the metric distribution.

Back to article page