Using Machine Learning for Earthquakes and Quarry Blasts Discrimination

The effects of explosions and other manmade seismic sources pose a threat to humanity. One of the most pressing issues currently confronting seismologists is contamination of seismicity catalogs. In order to distinguish tectonic from non-tectonic occurrences, an automated control system must be developed, and since detecting quarry blasts (QBs) is the initial and always tough stage, this is an absolute necessity. The need to locate and eliminate the man-made seismic disturbances has increased dramatically. In order to aid in precise seismic hazard identification and improve the planning of future urban developments, early treatments and cleaning of contaminated seismicity catalogs are necessary. Machine learning (ML) methods have allowed for greater precision in identifying synthetic seismic sources. Distinguishing between QBs and natural earthquakes is currently the focus of numerous methodologies, ML techniques, and varied processes, such as knowledge discovery. In order for intelligent systems to learn from repeated encounters and spot and identify patterns in a dataset, ML techniques provide a variety of probabilistic and statistical methods. The purpose of this research is to develop an algorithm that can identify QBs inside seismicity databases automatically. To be more specific, we use classical and ensemble ML classifiers to categorize reports of seismic activity. In order to improve performance, the suggested technique makes use only three features (Latitude, Longitude, and Magnitude). The accuracy of the proposed scheme is examined by R2, F1-score, MCC score, kappa score, elapsed time, learning curve, and confusion matrix. The proposed mode has demonstrated the superior performance as compared to the benchmarks with a testing accuracy of 97.21 %.