Automatic Identification and Analysis of Natural Earthquake and Artificial Blasting in Inner Mongolia Region Seismic Network Based on Support Vector Machine
Abstract We use natural earthquakes and blasting events in Inner Mongolia and its surrounding areas (96°-126°E, 36°-54°N) from 2016 to 2021. Firstly, we use db7, sym6 and rbio1.5 wavelet basis functions to decompose the event waveforms into discrete wavelet, static wavelet and wavelet packet, and then extract three characteristic parameters of energy ratio, Shannon entropy and energy entropy. 288 groups of experiments are conducted by random combination of different wavelet decomposition methods, kernel function, support vector machine and eigenvalue. The results show that the highest recognition rate of “DWT+υ-SVC+db7+linear kernel+energy ratio+ Shannon entropy+energy entropy” is 95%, which indicates this method is more suitable for Inner Mongolia. It can provide a reliable reference for the identification of natural earthquake and blasting events in seismic network.
WANG Lujun,YIN Zhanjun,ZHAI Hao et al. Automatic Identification and Analysis of Natural Earthquake and Artificial Blasting in Inner Mongolia Region Seismic Network Based on Support Vector Machine[J]. jgg, 2022, 42(3): 326-330.
WANG Lujun,YIN Zhanjun,ZHAI Hao et al. Automatic Identification and Analysis of Natural Earthquake and Artificial Blasting in Inner Mongolia Region Seismic Network Based on Support Vector Machine[J]. jgg, 2022, 42(3): 326-330.
