Citation

Saidani O, Manoharan RR, Naje AS, Mishra R, Subburaj A, S M, Sikkandar MY, Sundaram SG, Rajan R, Sengan S. Earth Sci. Inform. 2022; 15(3): 1527-1544.

Copyright

(Copyright © 2022, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s12145-021-00754-8

PMID

unavailable

Abstract

Earthquake Prediction has become a field of seismology concerned with the specification of time, location, and magnitude of earthquakes to take preventive measures that could help in worst-case scenarios, i.e., destruction of homes and lives. Short Term Earthquakes depends on anomalous events known as precursors that occur before an Earthquake. Precursors are considered a warning before an earthquake. This prediction system uses Earth's Electric Field Signal (EEFS) from Athens, Pyrgos, and Hios. Some examples are Ground Uplifting, Tilting, Emission of Radon Gas, Radio Waves, Magnetic Waves, and Earth's Electric. The EEFS can be considered a precursor to estimate the magnitude and time of a possible earthquake. A minute-by-minute reading of the Earth's Electric Field was taken, and few computations, models like ANN, SVM-ANN, and SVM-KNN, were applied. Few other models were created to estimate the time and magnitude of the earthquake. To discover an ideal model, results are compared without any constraints such as overfitting. The paper presents directions for estimating time and is directing for researchers to analyze in multiple dimensions.

Language: en

Keywords

Artificial Neural Networks; Earthquake Prediction; Ensemble Techniques; Magnitude Estimation; Time Estimation