Current Volume 8
In particular, convolutional neural networks (CNNs) have become useful and flexible instruments in approaching seismic data flows and have proved their capacity to increase seismic analysis and forecast seismic occurrences. Although it is true that CNNs are a novel area of study and are not commonly used in earthquake prediction, in this paper, we discuss its use in earthquake prediction with the capability of giving the algorithm seismic data as raw inputs and the identification of patterns which are otherwise undetectable. CNNs have been successfully applied in the analysis of real-time event detection and classification, with differentiation in earthquake magnitude and depth, aftershocks, and ground-motion prediction. However, there are several limitations, including data deficit, difficulty in cross-geological area model deployment and the limitations of DL algorithms in interpretation. Concerning CNNs combined with other ML algorithms and future trends of CNN-based earthquake prediction systems, this paper also presents the following developments: hybrid models, transferring learning and expanding the network of seismological stations throughout the globe. CNN frameworks of using models can greatly improve earthquake hazard risk reduction frameworks, which supplement new early warning systems and the effectiveness of disaster preparedness.
Convolutional Neural Networks, Earthquake Prediction, Seismic Data, Deep Learning, Earthquake Early Warning, Ground Motion Prediction, Aftershock Forecasting.
IRE Journals:
Mohit Jain , Adit Shah
"Machine Learning with Convolutional Neural Networks (CNNs) in Seismology for Earthquake Prediction" Iconic Research And Engineering Journals Volume 5 Issue 8 2022 Page 389-398
IEEE:
Mohit Jain , Adit Shah
"Machine Learning with Convolutional Neural Networks (CNNs) in Seismology for Earthquake Prediction" Iconic Research And Engineering Journals, 5(8)