Abstract—The distinction between earthquake and explosion signals is an essential issue in seismic signal analysis. We propose a machine learning model of decision tree (DT) applied to discriminate between earthquakes and explosions. The amplitudes of the P-wave and the S-wave are selected as feature vectors and built into the database. Classification and regression trees (CART) algorithm is used in our method, which is built through a greedy approach by the Gini impurity. The performance of the DT model using the CART algorithm is evaluated with the ROC curve. The results show the advantages of DT according to various evaluation indexes based on confusion matrix, and demonstrate that DT is efficient in seismic signal discrimination due to the nonparametric model characteristics of DT model.
Index Terms—Decision tree, seismic wave, classification and regression tree, receiver operating characteristic.
The authors are with the Department of Electrical and Computer Engineering, Sungkyunkwan University, Suwon, Korea (e-mail: lhw2065@skku.edu, naman2001@skku.edu, khyou@skku.edu).
[PDF]
Cite: Hyukwoo Lee, Kyunghyun Lee, and Kwanho You, "Seismic Discrimination between Earthquakes and Explosions Using CART Algorithm," International Journal of Materials, Mechanics and Manufacturing vol. 7, no. 1, pp. 51-54, 2019.
![[Click]](http://www.ijmmm.org/img/ijmmm_v7n2_inspec%28IET%29_indexscreenshot.jpg){kind=link}