In this study, a computer‐aided methodology is proposed to estimate the earthquake magnitude based on fault parameters. So far, log–linear regression equations are separately employed for each fault parameter. However, this can lead to inconsistent magnitude predictions because nonlinear parameter correlations are ignored and those parametric functions cannot take into account potential deviations from log–linear scaling. To address the aforementioned deficiencies, we employ Artificial Neural Network (ANN) to estimate the magnitude of earthquakes simultaneously using all available fault parameters such as rupture length and width, thereby excluding the chances of inconsistent estimations. Our evaluation of M5 earthquakes shows that the predictions from the proposed methodology outperform the regression‐equation‐based predictions in terms of mean absolute error and root mean square error. Furthermore, the pictorial view of the performance also demonstrates the strength of ANN to identify and reproduce, without any initial assumption, systematic deviations from the log–linear scaling of earthquake magnitudes as a function of the fault parameters.

You do not currently have access to this article.