Abstract
Earthquake precursor data have been used as an important basis for earthquake prediction. In this study, a recurrent neural network (RNN) architecture with long short-term memory (LSTM) units is utilized to develop a predictive model for normal data. Furthermore, the prediction errors from the predictive models are used to indicate normal or abnormal behavior. An additional advantage of using the LSTM networks is that the earthquake precursor data can be directly fed into the network without any elaborate preprocessing as required by other approaches. Furthermore, no prior information on abnormal data is needed by these networks as they are trained only using normal data. Experiments using three groups of real data were conducted to compare the anomaly detection results of the proposed method with those of manual recognition. The comparison results indicated that the proposed LSTM network achieves promising results and is viable for detecting anomalies in earthquake precursor data.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Bengio, Y., Simard, P., and Frasconi, P., 1994, Learning long-term dependencies with gradient descent is diffi cult: IEEE Transactions on Neural Networks, 5(2), 157–166.
Chung, J., Gulcehre, C., Cho, K., and Bengio, Y., 2014, Empirical evaluation ofgated recurrent neural networks on sequence modeling.
Dey, R., and Salemt, F.M., 2017, Gate-variants of gated recurrent unit (GRU) neural networks: IEEE 60th International Midwest Symposium on Circuits and Systems (MWSCAS), 1597–1600.
Frinken, V., and Uchida, S., 2015, Deep BLSTM neural networks for unconstrained continuous handwritten text recognition: 3th International Conference on Document Analysis and Recognition (ICDAR), IEEE, 911–915.
Gers, F., Schmidhuber, J., and Cummins, F., 2000, Learning to forget: continual prediction with LSTM: Neural Computation, 12(10), 2451–2471.
Graves, A., Mohamed, A.-r., and Hinton, G., 2013, Speech recognition with deep recurrent neural networks: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE, 6645–6649.
Greff, K., Srivastava, R.K., Koutn J., Steunebrink, B.R., and Schmidhuber, J., 2017, LSTM: A search space odyssey: IEEE Transactions on Neural Networks Learning Systems, 28(10), 2222–2232.
Grossmann, A., and Morlet, J., 1984, Decomposition of hardy functions into square integrable wavelets of constant shape: SIAM Journal on Mathematical Analysis, 15(4), 723–736.
Gupta, M., Gao, J., Aggarwal, C.C., and Han, J., 2014, Outlier detection for temporal data: A survey: IEEE Transactions on Knowledge Data Engineering, 26(9), 2250–2267.
Han, P., Hattori, K., Xu, G., Ashida, R., Chen, C.- H., Febriani, F., and Yamaguchi, H., 2015, Further investigations of geomagnetic diurnal variations associated with the 2011 off the Pacific coast of Tohoku earthquake (Mw 9.0): Journal of Asian Earth Sciences, 114, 321–326.
Hermans, M., and Schrauwen, B., 2013, Training and analysing deep recurrent neural networks: Advances in neural information processing systems, 190–198.
Hochreiter, S., Bengio, Y., Frasconi, P., and Schmidhuber, J., 2001, Gradient fl ow in recurrent nets: the difficulty of learning long-term dependencies: A field guide to dynamical recurrent neural network, Wiley-IEEE Press, USA, 1–15.
Hochreiter, S., and Schmidhuber, J., 1997, Long shortterm memory: Neural Computation, 9(8), 1735–1780.
Huang, F., Li, M., Ma, Y., Han, Y., Tian, L., Yan, W. and Li, X., 2017, Studies on earthquake precursors in China: A review for recent 50 years: Geodesy and Geodynamics, 8(1), 1–12.
Huang, N.E., Shen, Z., Long, S.R., Wu, M.C., Shih, H.H., Zheng, Q., Yen, N.-C., Tung, C.C., and Liu, H.H., 1998, The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis: Proceedings of the Royal Society of London. Series A: Mathematical, Physical and 10 Anomaly detection of earthquake precursor data Engineering Sciences, 454(1971), 903–995.
Huang, Q., 2011, Retrospective investigation of geophysical data possibly associated with the Ms8. 0 Wenchuan earthquake in Sichuan, China: Journal of Asian Earth Sciences, 41(4-5), 421–427.
LeCun, Y., Bengio, Y., and Hinton, G., 2015, Deep learning: Nature, 521(7553), 436–444.
Lu, K., Hou, M., Jiang, Z., Wang, Q., Sun, G., and Liu, J., 2018, Understanding earthquake from the granular physics point of view Causes of earthquake, earthquake precursors and predictions: International Journal of Modern Physics B, 32(07), 1–39.
Masci, F., and Thomas, J.N., 2015, Ultra low frequency (ULF) electromagnetic anomalies associated with large earthquakes in Java Island, Indonesia by using wavelet transform and detrended fl uctuation analysis: Natural Hazards and Earth System Sciences, 15(12), 2697–2701.
Ouyang, Z., Zhang, H., Fu, Z., Gou, B., and Jiang, W., 2009, Abnormal phenomena recorded by several earthquake precursor observation instruments before the Ms 8.0 Wenchuan, Sichuan earthquake: Acta Geologica Sinica-English Edition, 83(4), 834–844.
Pouyanfar, S., Sadiq, S., Yan, Y., Tian, H., Tao, Y., Reyes, M.P., Shyu, M.-L., Chen, S.-C., and Iyengar, S., 2018, A survey on deep learning: algorithms, techniques, and applications: ACM Computing Surveys, 51(5), 92.
Sak, H., Senior, A., and Beaufays, F., 2014, Long short-term memory recurrent neural network architectures for large scale acoustic modeling: 15th Annual Conference of the International Speech Communication Association, 338–342.
Shah, M., and Jin, S., 2018, Pre-seismic ionospheric anomalies of the 2013 Mw = 7.7 Pakistan earthquake from GPS and COSMIC observations: Geodesy and Geodynamics, 9(5), 378–387.
Shirokov, V., Firstov, P., Makarov, E., and Stepanov, I., 2015, An Approach to the short-and long-term forecasting of strong earthquakes: A case study of the Mw = 9.0 Tohoku earthquake, Japan, March 11, 2011: Seismic Instruments, 51(3), 229–241.
Soltau, H., Liao, H., and Sak, H., 2016, Neural speech recognizer: Acoustic-to-word LSTM model for large vocabulary speech recognition: arXiv preprint arXiv:1610.09975
Tsai, Y.B., Liu, J.Y., Shin, T.C., Yen, H.Y., and Chen, C.H., 2018, Multidisciplinary earthquake precursor studies in Taiwan: A review and future prospects: Preearthquake processes: A Multidisciplinary Approach to Earthquake Prediction Studies, John Wiley & Sons, Inc., USA, 41–65.
Wang, J., Chen, X., Kang, B., and Li, X., 2017, Groundwater micro-dynamic extraction and its response to seismic activity based on EMD: Ninth International Conference on Digital Image Processing (ICDIP 2017), SPIE, 1–5.
Wang, K., Chen, Q.-F., Sun, S., and Wang, A., 2006, Predicting the 1975 Haicheng earthquake: Bulletin of the Seismological Society of America, 96(3), 757–795.
Werbos, P.J., 1990, Backpropagation through time: what it does and how to do it: Proceedings of the IEEE, 78(10), 1550–1560.
Williams, R.J., and Zipser, D., 1989, A learning algorithm for continually running fully recurrent neural networks: Neural Computation, 1(2), 270–280.
Wu, Z., Ma, T., Jiang, H., and Jiang, C., 2013, Multiscale seismic hazard and risk in the China mainland with implication for the preparedness, mitigation, and management of earthquake disasters: An overview: International Journal of Disaster Risk Reduction, 4, 21–33.
Yang, D.-H., Yuan, J., Wang, X.-Y., Shen, X.-H., Teng, H.-T., Li, W.-J., Tan, Q., and Wei, Q., 2017, Identification of multi-anomalies of precursory deformation data: Chinese Journal of Geophys-Chinese Edition (in Chinese), 60(12), 4623–4632.
Yao, Y., Chen, P., Wu, H., Zhang, S., and Peng, W., 2012, Analysis of ionospheric anomalies before the 2011 Mw 9.0 Japan earthquake: Chinese Science Bulletin, 57(5), 500–510.
Ye, Q., Singh, R.P., He, A., Ji, S., and Liu, C., 2015, Characteristic behavior of water radon associated with Wenchuan and Lushan earthquakes along Longmenshan fault: Radiation Measurements, 76, 44–53.
Zhang, W., Huang, W., Li, L., Liu, W., and Li, F., 2016, High resolution strain sensor for earthquake precursor observation and earthquake monitoring: Sixth European Workshop on Optical Fibre Sensors (EWOFS’2016), SPIE, 1–5.
Acknowledgments
The authors appreciate the helpful discussion and technical support provided by the DDM Research Group of the University of Miami.
Author information
Authors and Affiliations
Corresponding author
Additional information
This work was supported by the Science for Earthquake Resilience of China (No. XH18027),Research and Development of Comprehensive Geophysical Field Observing Instrument in Mainland China(No. Y201703) and Research Fund Project of Shandong Earthquake Agency(Nos. JJ1505Y and JJ1602).
Yin Cai is a senior engineer in Shandong Earthquake Agency and graduated from Dalian University of Technology in 2007, with a master’s degree in software engineering. He is a Ph.D. candidate in Institute of Geophysics, China Earthquake Administration. His researches primarily focus on the application of deep learning methods and cloud computing technology in geoscience.
Rights and permissions
About this article
Cite this article
Cai, Y., Shyu, ML., Tu, YX. et al. Anomaly detection of earthquake precursor data using long short-term memory networks. Appl. Geophys. 16, 257–266 (2019). https://doi.org/10.1007/s11770-019-0774-1
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11770-019-0774-1