Proposed ConvBiLSTM-Net Model for Enhancing Earthquake Prediction Performance Using Spatiotemporal Features

Authors

DOI:

https://doi.org/10.31436/iiumej.v26i3.3634

Keywords:

Deep Learning, Earthquake Prediction, Kernel Density Estimation (KDE), Temporal and Spatial Features

Abstract

Accurate earthquake prediction remains a significant challenge due to the complex spatiotemporal dependencies inherent in seismic events. To address this issue, the present study proposes ConvBiLSTM-Net. This hybrid deep learning model combines Convolutional Neural Networks (CNNs) for spatial feature extraction with Bidirectional Long Short-Term Memory (BiLSTM) networks for temporal sequence modeling. The model integrates historical earthquake data with spatial information in the form of fault density (FD), derived using Kernel Density Estimation (KDE). The KDE bandwidth is optimized using the Bivariate Local Indicator of Spatial Association (LISA) method to enhance spatial adaptivity. The dataset comprises earthquake records from the USGS catalog (1974–2023) and active fault data compiled in the 2017 Indonesian Earthquake Source and Hazard Map, published by the National Earthquake Study Center (PuSGeN). ConvBiLSTM-Net is evaluated under short-term and medium-term prediction scenarios, targeting earthquake magnitude, depth, and epicenter coordinates (latitude and longitude), using standard performance metrics such as accuracy, F1 score, root mean square error (RMSE), mean absolute error (MAE), and the coefficient of determination (R²). In the short-term scenario, the model achieves average improvements of 9.31% in R², 3.41% in accuracy, and 6.06% in F1 score, while reducing RMSE by 10.63% and MAE by 12.40% across magnitude, depth, and latitude predictions. For longitude, R², accuracy, and F1 score also improve by 10.88%, 11.76%, and 17.54%, respectively, although RMSE and MAE increase by 13.09% and 20.74%, indicating a trade-off between enhanced pattern recognition and higher absolute error. Under the medium-term scenario, the model demonstrates average improvements of 7.49% in R², 3.39% in accuracy, and 7.06% in F1 score, while reducing RMSE and MAE by 6.22% and 17.72%, respectively, for magnitude, depth, and latitude predictions. For longitude, R², accuracy, and F1 score improve by 12.50%, 2.48%, and 1.60%, respectively, though RMSE and MAE increase by 37.31% and 37.01%, again highlighting a trade-off between better pattern recognition and increased absolute error in this dimension. These findings demonstrate that ConvBiLSTM-Net, engineered to integrate spatial and temporal features, is a robust and adaptive architecture for enhancing earthquake prediction performance. Its spatiotemporal modeling approach yields consistently high accuracy and stability across forecasting horizons, particularly in predicting earthquake epicenters. Despite minor trade-offs in absolute error for longitude predictions, the overall performance improvements affirm its potential as a reliable tool for seismic hazard assessment and disaster risk mitigation.

ABSTRAK: Ramalan gempa bumi yang tepat kekal sebagai satu cabaran utama disebabkan oleh kebergantungan spatiotemporal yang kompleks dalam kejadian seismik. Bagi menangani isu ini, kajian ini mencadangkan ConvBiLSTM-Net, iaitu sebuah model hibrid pembelajaran mendalam yang menggabungkan Rangkaian Neural Konvolusional (CNN) dan Memori Jangka Pendek Dwi Arah (BiLSTM), bagi tujuan pengekstrakan ciri spatial dan pemodelan jujukan temporal, masing-masing.  Model ini menggabungkan data sejarah gempa bumi dengan maklumat spatial dalam bentuk ketumpatan sesar (fault density, FD), yang diperoleh melalui Kaedah Anggaran Ketumpatan Kernel (Kernel Density Estimation, KDE). Lebar jalur KDE dioptimumkan menggunakan kaedah Bivariate Local Indicator of Spatial Association (LISA) bagi meningkatkan kepekaan spatial. Set data kajian merangkumi rekod gempa bumi daripada katalog USGS (1974–2023) serta data sesar aktif yang disusun dalam Peta Sumber dan Bahaya Gempa Indonesia 2017, terbitan Pusat Kajian Gempa Nasional (PuSGeN). Model ConvBiLSTM-Net ini dinilai dalam dua senario ramalan—jangka pendek dan jangka sederhana—bagi parameter magnitud, kedalaman, serta koordinat pusat gempa (latitud dan longitud), dengan menggunakan metrik standard seperti ketepatan, skor F1, RMSE, MAE dan pekali penentuan (R²). Malalui senario jangka pendek, model mencatatkan purata peningkatan sebanyak 9.31% pada R², 3.41% pada ketepatan, dan 6.06% pada skor F1; serta pengurangan RMSE sebanyak 10.63% dan MAE sebanyak 12.40% merentas ramalan magnitud, kedalaman, dan latitud. Bagi dimensi longitud, R², ketepatan, dan skor F1 turut meningkat sebanyak 10.88%, 11.76%, dan 17.54% masing-masing; namun begitu, RMSE dan MAE meningkat sebanyak 13.09% dan 20.74%, menunjukkan kompromi antara pengecaman corak yang lebih baik dan ralat mutlak yang lebih tinggi. Manakala senario jangka sederhana, model mencatatkan purata peningkatan sebanyak 7.49% pada R², 3.39% pada ketepatan, dan 7.06% pada skor F1; serta pengurangan RMSE sebanyak 6.22% dan MAE sebanyak 17.72% merentas tugas ramalan magnitud, kedalaman, dan latitud. Bagi dimensi longitud, peningkatan masing-masing dicatatkan pada R² (12.50%), ketepatan (2.48%), dan skor F1 (1.60%), tetapi RMSE dan MAE meningkat secara ketara sebanyak 37.31% dan 37.01%, menunjukkan kompromi antara pengecaman corak yang lebih kukuh dan ralat mutlak yang lebih besar pada dimensi ini. Dapatan kajian ini membuktikan bahawa ConvBiLSTM?Net, yang direka bentuk bagi menggabungkan ciri-ciri spatial dan temporal, merupakan satu seni bina model yang teguh dan adaptif dalam meningkatkan prestasi ramalan gempa bumi. Pemodelan spatiotemporal bersepadu yang digunakan menghasilkan tahap ketepatan dan kestabilan yang tinggi secara konsisten merentasi pelbagai ufuk ramalan, terutamanya dalam menentukan lokasi pusat gempa bumi. Walaupun terdapat sedikit kekurangan dalam nilai ralat mutlak bagi ramalan longitud, peningkatan prestasi secara keseluruhan mengesahkan nilainya sebagai alat yang boleh dipercayai dalam penilaian bahaya seismik dan pengurangan risiko bencana.

Downloads

Download data is not yet available.

Metrics

Metrics Loading ...

Author Biographies

Ari Fadli, Universitas Gadjah Mada

 

 Department of Computer Science and Electronics, Universitas Gadjah Mada, Yogyakarta, Indonesia 

 Department of Electrical Engineering, Universitas Jenderal Soedirman, Purwokerto, Indonesia 

Agfianto Eko Putra, Universitas Gadjah Mada

 

 Department of Computer Science and Electronics, Universitas Gadjah Mada, Yogyakarta 

Wiwit Suryanto, Universitas Gadjah Mada

 

 Department of Physics, Universitas Gadjah Mada, Yogyakarta, Indonesia 

References

Alarifi, A.S.N., Alarifi, N.S.N., Al-Humidan, S.: Earthquakes magnitude predication using artificial neural network in northern Red Sea area. J King Saud Univ Sci. 24, 301–313 (2012). https://doi.org/10.1016/j.jksus.2011.05.002.

Berhich, A., Belouadha, F.-Z., Kabbaj, M.I.: An attention-based LSTM network for large earthquake prediction. Soil Dynamics and Earthquake Engineering. 165, 107663 (2023). https://doi.org/10.1016/j.soildyn.2022.107663.

Shidik, G.F., Pramunendar, R.A., Kusuma, E.J., Saraswati, G.W., Winarsih, N.A.S., Rohman, M.S., Saputra, F.O., Naufal, M., Andono, P.N.: LUTanh Activation Function to Optimize BI-LSTM in Earthquake Forecasting. International Journal of Intelligent Engineering and Systems. 17, 572 – 583 (2024). https://doi.org/10.22266/ijies2024.0229.48.

Maharani, E.: Infografis Kerusakan Ekonomi Akibat Gempa Besar di Dunia, https://visual.republika.co.id/berita/rrsyo1335/infografis-kerusakan-ekonomi-akibat-gempa-besar-di-dunia, last accessed 2023/10/01.

UYEDA, S.: On Earthquake Prediction in Japan. Proceedings of the Japan Academy, Series B. 89, 391–400 (2013). https://doi.org/10.2183/pjab.89.391.

HAYAKAWA, M., SCHEKOTOV, A., POTIRAKIS, S., EFTAXIAS, K.: Criticality features in ULF magnetic fields prior to the 2011 T?hoku earthquake. Proceedings of the Japan Academy, Series B. 91, 25–30 (2015). https://doi.org/10.2183/pjab.91.25.

Alimoradi, H., Rahimi, H., Ovaisi Moaakhar, M.: Investigation of changes in the cumulative number of magnetic anomalies before and after earthquakes using satellite data. Advances in Space Research. 75, 895–907 (2025). https://doi.org/10.1016/j.asr.2024.09.012.

Wang, Q., Guo, Y., Yu, L., Li, P.: Earthquake Prediction Based on Spatio-Temporal Data Mining: An LSTM Network Approach. IEEE Trans Emerg Top Comput. 8, 148–158 (2017). https://doi.org/10.1109/TETC.2017.2699169.

Kagan, Y.Y., Jackson, D.D.: Long-Term Earthquake Clustering. Geophys J Int. 104, 117–134 (1991). https://doi.org/10.1111/j.1365-246X.1991.tb02498.x.

Kagan, Y.Y., Jackson, D.D.: Long?term probabilistic forecasting of earthquakes. J Geophys Res Solid Earth. 99, 13685–13700 (1994). https://doi.org/10.1029/94JB00500.

Banna, Md.H. Al, Taher, K.A., Kaiser, M.S., Mahmud, M., Rahman, Md.S., Hosen, A.S.M.S., Cho, G.H.: Application of Artificial Intelligence in Predicting Earthquakes: State-of-the-Art and Future Challenges. IEEE Access. 8, 192880–192923 (2020). https://doi.org/10.1109/ACCESS.2020.3029859.

Mojtahedi, F.F., Yousefpour, N., Chow, S.H., Cassidy, M.: Deep Learning for Time Series Forecasting: Review and Applications in Geotechnics and Geosciences. Archives of Computational Methods in Engineering. (2025). https://doi.org/10.1007/s11831-025-10244-5.

Majumder, R.Q.: Machine Learning for Predictive Analytics: Trends and Future Directions. SSRN Electronic Journal. (2025). https://doi.org/10.2139/ssrn.5267273.

Jamarani, A., Haddadi, S., Sarvizadeh, R., Haghi Kashani, M., Akbari, M., Moradi, S.: Big data and predictive analytics: A systematic review of applications. Artif Intell Rev. 57, 176 (2024). https://doi.org/10.1007/s10462-024-10811-5.

Huang, C.-J., Kuo, P.-H.: A Short-Term Wind Speed Forecasting Model by Using Artificial Neural Networks with Stochastic Optimization for Renewable Energy Systems. Energies (Basel). 11, 2777 (2018). https://doi.org/10.3390/en11102777.

Peng, L., Wang, L., Ai, X.-Y., Zeng, Y.-R.: Forecasting Tourist Arrivals via Random Forest and Long Short-term Memory. Cognit Comput. 13, 125–138 (2021). https://doi.org/10.1007/s12559-020-09747-z.

Li, X., Wu, P.: Stock Price Prediction Incorporating Market Style Clustering. Cognit Comput. 14, 149–166 (2022). https://doi.org/10.1007/s12559-021-09820-1.

Wang, X., Nguyen, K., Nguyen, B.P.: Churn Prediction using Ensemble Learning. In: Proceedings of the 4th International Conference on Machine Learning and Soft Computing. pp. 56–60. ACM, New York, NY, USA (2020). https://doi.org/10.1145/3380688.3380710.

Asim, K.M., Martínez-Álvarez, F., Basit, A., Iqbal, T.: Earthquake magnitude prediction in Hindukush region using machine learning techniques. Natural Hazards. 85, 471–486 (2017). https://doi.org/10.1007/s11069-016-2579-3.

Asim, K.M., Idris, A., Iqbal, T., Martínez-Álvarez, F.: Earthquake prediction model using support vector regressor and hybrid neural networks. PLoS One. 13, e0199004 (2018). https://doi.org/10.1371/journal.pone.0199004.

Yaghmaei-Sabegh, S., Tsang, H.-H.: A new site classification approach based on neural networks. Soil Dynamics and Earthquake Engineering. 31, 974–981 (2011). https://doi.org/10.1016/j.soildyn.2011.03.004.

Asim, K.M., Moustafa, S.SR., Niaz, I.A., Elawadi, E.A., Iqbal, T., Martínez-Álvarez, F.: Seismicity analysis and machine learning models for short-term low magnitude seismic activity predictions in Cyprus. Soil Dynamics and Earthquake Engineering. 130, 105932 (2020). https://doi.org/10.1016/j.soildyn.2019.105932.

Jain, R., Nayyar, A., Arora, S., Gupta, A.: A comprehensive analysis and prediction of earthquake magnitude based on position and depth parameters using machine and deep learning models. Multimed Tools Appl. 80, 28419–28438 (2021). https://doi.org/10.1007/s11042-021-11001-z.

Gonzalez, J., Yu, W., Telesca, L.: Gated Recurrent Units Based Recurrent Neural Network for Forecasting the Characteristics of the Next Earthquake. Cybern Syst. 53, 209–222 (2022). https://doi.org/10.1080/01969722.2021.1981637.

Kavianpour, P., Kavianpour, M., Ramezani, A.: Deep Multi-scale Dilated Convolution Neural Network with Attention Mechanism: A Novel Method for Earthquake Magnitude Classification. In: 2022 8th Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS). pp. 1–6 (2022). https://doi.org/10.1109/ICSPIS56952.2022.10043978.

Cai, Y., Shyu, M.-L., Tu, Y.-X., Teng, Y.-T., Hu, X.-X.: Anomaly detection of earthquake precursor data using long short-term memory networks. Applied Geophysics. 16, 257–266 (2019). https://doi.org/10.1007/s11770-019-0774-1.

Fabregas, A.C., Arellano, P.B. V., Pinili, A.N.D.: Long-Short Term Memory (LSTM) Networks with Time Series and Spatio-Temporal Approaches Applied in Forecasting Earthquakes in the Philippines. In: Proceedings of the 4th International Conference on Natural Language Processing and Information Retrieval. pp. 188–193. ACM, New York, NY, USA (2020). https://doi.org/10.1145/3443279.3443288.

Berhich, A., Belouadha, F.-Z., Kabbaj, M.I.: An attention-based LSTM network for large earthquake prediction. Soil Dynamics and Earthquake Engineering. 165, 107663 (2023). https://doi.org/10.1016/j.soildyn.2022.107663.

Kail, R., Burnaev, E., Zaytsev, A.: Recurrent Convolutional Neural Networks Help to Predict Location of Earthquakes. IEEE Geoscience and Remote Sensing Letters. 19, 1–5 (2022). https://doi.org/10.1109/LGRS.2021.3107998.

Bhandarkar, T., K, V., Satish, N., Sridhar, S., Sivakumar, R., Ghosh, S.: Earthquake trend prediction using long short-term memory RNN. International Journal of Electrical and Computer Engineering (IJECE). 9, 1304 (2019). https://doi.org/10.11591/ijece.v9i2.pp1304-1312.

Kavianpour, P., Kavianpour, M., Jahani, E., Ramezani, A.: A CNN-BiLSTM model with attention mechanism for earthquake prediction. J Supercomput. 80, 2913–2913 (2024). https://doi.org/10.1007/s11227-023-05497-5.

Kavianpour, P., Kavianpour, M., Jahani, E., Ramezani, A.: Earthquake Magnitude Prediction using Spatia-Temporal Features Learning Based on Hybrid CNN-BiLSTM Model. In: Proceedings - 2021 7th International Conference on Signal Processing and Intelligent Systems, ICSPIS 2021. Institute of Electrical and Electronics Engineers Inc. (2021). https://doi.org/10.1109/ICSPIS54653.2021.9729358.

Geng, Y., Su, L., Jia, Y., Han, C.: Seismic Events Prediction Using Deep Temporal Convolution Networks. Journal of Electrical and Computer Engineering. 2019, 1–14 (2019). https://doi.org/10.1155/2019/7343784.

Yousefzadeh, M., Hosseini, S.A., Farnaghi, M.: Spatiotemporally explicit earthquake prediction using deep neural network. Soil Dynamics and Earthquake Engineering. 144, 106663 (2021). https://doi.org/10.1016/j.soildyn.2021.106663.

Hetényi, G., Epard, J.-L., Colavitti, L., Hirzel, A.H., Kiss, D., Petri, B., Scarponi, M., Schmalholz, S.M., Subedi, S.: Spatial relation of surface faults and crustal seismicity: a first comparison in the region of Switzerland. Acta Geodaetica et Geophysica. 53, 439–461 (2018). https://doi.org/10.1007/s40328-018-0229-9.

Matsuda, T.: Estimation of Future Destructive Earthquakes from Active Faults on Land in Japan. Journal of Physics of the Earth. 25, S251–S260 (1977). https://doi.org/10.4294/jpe1952.25.Supplement_S251.

Matsuda, T.: Active Faults and Damaging Earthquakes in Japan-Macroseismic Zoning and Precaution Fault Zones. (2013). https://doi.org/10.1029/ME004p0279.

Sharma, A., Ahuja, A., Devi, S., Pasari, S.: Use of Spatio-temporal Features for Earthquake Forecasting of imbalanced Data. In: 2022 International Conference on Intelligent Innovations in Engineering and Technology (ICIIET). pp. 178–182. IEEE (2022). https://doi.org/10.1109/ICIIET55458.2022.9967687.

Song, W., Wang, C., Chen, W., Zhang, X., Li, H., Li, J.: Unlocking the spatial heterogeneous relationship between Per Capita GDP and nearby air quality using bivariate local indicator of spatial association. Resour Conserv Recycl. 160, 104880 (2020). https://doi.org/10.1016/j.resconrec.2020.104880.

Phan, H., Andreotti, F., Cooray, N., Chen, O.Y., De Vos, M.: DNN Filter Bank Improves 1-Max Pooling CNN for Single-Channel EEG Automatic Sleep Stage Classification. In: 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). pp. 453–456. IEEE (2018). https://doi.org/10.1109/EMBC.2018.8512286.

Qiao, M., Yan, S., Tang, X., Xu, C.: Deep Convolutional and LSTM Recurrent Neural Networks for Rolling Bearing Fault Diagnosis Under Strong Noises and Variable Loads. IEEE Access. 8, 66257–66269 (2020). https://doi.org/10.1109/ACCESS.2020.2985617.

PuSGeN: Peta Sumber Daya dan Bahaya Gempa Indonesia Tahun 2017. Puslitbang PUPR, Indonesia (2017).

Adeli, H., Panakkat, A.: A probabilistic neural network for earthquake magnitude prediction. Neural Networks. 22, 1018–1024 (2009). https://doi.org/10.1016/j.neunet.2009.05.003.

Reyes, J., Morales-Esteban, A., Martínez-Álvarez, F.: Neural networks to predict earthquakes in Chile. Appl Soft Comput. 13, 1314–1328 (2013). https://doi.org/10.1016/j.asoc.2012.10.014.

Anselin, L.: Local Indicators of Spatial Association, LISA. Geogr Anal. 27, 93–115 (1995). https://doi.org/10.1111/j.1538-4632.1995.tb00338.x.

Lee, J., Wong, D.W.S.: Statistical Analysis with ArcView GIS. John Wiley & Sons, New York (2001).

Mathur, M.: Spatial autocorrelation analysis in plant population: An overview. Journal of Applied and Natural Science. 7, 501–513 (2015). https://doi.org/10.31018/jans.v7i1.639.

Gao, Y., Cheng, J., Meng, H., Liu, Y.: Measuring spatio-temporal autocorrelation in time series data of collective human mobility. Geo-spatial Information Science. 22, 166–173 (2019). https://doi.org/10.1080/10095020.2019.1643609.

Downloads

Published

2025-09-09

How to Cite

Fadli, A., Priyambodo, T. K., Putra, A. E., & Suryanto, W. (2025). Proposed ConvBiLSTM-Net Model for Enhancing Earthquake Prediction Performance Using Spatiotemporal Features. IIUM Engineering Journal, 26(3), 238–259. https://doi.org/10.31436/iiumej.v26i3.3634

Issue

Section

Electrical, Computer and Communications Engineering

Most read articles by the same author(s)