Investigating the Effect of Data Augmentation on the Intelligentization of Environmental Hazard Studies - Case Study: Real-Time Calculation of Earthquake Magnitudes in Early Warning Systems

Subject Areas :

Rezvan Esmaely ¹ , Roohollah Kimiaefar ² , Alireza Hajian ³ , Khosro Soleimani ⁴ , Maryam Hodhodi ⁵

1 - Department of Physics, Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran.
2 - Department of Physics, Najafabad Branch, Islamic Azad University, Najafabad, Iran.
3 - Departement of Physics, Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
4 - Departement of Mathematics, Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
5 - Departement of Physics, Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran.

Received: 2024-08-26 Accepted : 2024-12-11 Published : 2025-02-03

Keywords: Intelligentization, earthquake magnitude, earthquake early warning system, powerful earth movement, data augmentation.,

Abstract :

Natural hazards, including earthquakes, are serious challenges for human societies. These phenomena not only threaten human lives, but also cause significant economic and social losses. Considering that some of these risks occur suddenly and unpredictably, it is necessary that the evaluation and optimal management of these risks should be placed on the agenda of policymakers and crisis management officials. Recent advances and achievements in the field of intelligent algorithms have been able to reduce the human ability to reduce the effects caused by the occurrence. Among the examples of these achievements are Earthquake Early Warning Systems (EEWS), which have been proposed as an effective tool in reducing damages and human casualties, and can issue an early warning to the population affected by the event by calculating the basic parameters of the earthquake and provide the authorities with the necessary information. Considering that one of the basic pillars of using learning algorithms is the existence of sufficient training data, in cases such as earthquake data where the number of available samples is insufficient, data augmentation techniques are used. In this research, the impact of data augmentation in the calculation of the earthquake has been investigated in the real-time and based on the data of the strong motion network. Based on the analysis of the data of more than three thousand earthquakes, it has been shown that the use of data augmentation has improved the generalization performance of the trained network over test data by 37%.

References:

امین طوسی، محمود (1401). انتقال سبک برای افزایش داده‌های آموزشی شبکه‌های کانولوشنی در شناسایی شعلۀ آتش. هوش محاسباتی در مهندسی برق. 13 (4)، 97-114. doi: 10.22108/isee.2021.124044.1490
ایمانی، مريم (1403). داده‌افزایی مبتنی بر شبکه‌ی مولد رقابتی برای آشکارسازی اهداف تصاویر ابر طیفی با استفاده از شبکه‌های عصبی کانولوشن. مجله بینایی ماشین و پردازش تصویر. مجله بینایی ماشین و پردازش تصویر. 13-24. https://jmvip.sinaweb.net/article_207014.html.
Hajian, A.; Nunnari, G. & Kimiaefar, R. (2023). Intelligent methods and motivations to use in volcanology and seismology. Intelligent Methods with Applications in Volcanology and Seismology. Cham: Springer International Publishing, pp. 1–17. doi: 10.1007/978-3-031-15432-4_1.
Cheng, Y.; Wang, X. & Zhao, H. (2019). Data augmentation techniques for improving disaster prediction models. International Journal of Disaster Risk Reduction. 34, 380-393.
Esmaeili, R. et al (2024). Performance enhancement of deep neural network using fusional data assimilation and divide-and-conquer approach; case study: Earthquake magnitude calculation. Neural Computing and Applications. https://doi.org/10.1007/s00521-024-10002-x.
Fernandes, P. M.; Silva, J. S. & Amraoui, M. (2019). Wildfire risk assessment using machine learning and data augmentation techniques. Forest Ecology and Management 432, 43-55.
Gonzalez, R.; Castro, J. & Vargas, A. (2022). Integrating intelligent algorithms with data augmentation for real-time disaster risk assessment. Computers & Geosciences 148, 104694.
King, G., & Zeng, L. (2001). Logistic Regression in Rare Events Data. Political Analysis. 9 (2), 137–163. DOI:10.1093/oxfordjournals.pan.a004868.
Li, J.; Liu, Y. & Wang, L. (2021). Enhancing disaster response with artificial intelligence: Opportunities and challenges. Disaster Science and Management. 30 (2), 160-176.
Lin, Y. et al (2022). Generating clusters of similar sizes by constrained balanced clustering. Applied Intelligence. 52, 5273–5289. https://doi.org/10.1007/s10489-021-02682-y.
Samadi, H.; Kimiaefar, R. & Hajian, A. (2022). Fast earthquake relocation using ANFIS neuro-fuzzy network trained based on the double difference method. Scientific Quarterly Journal of Geosciences. 32 (3), 93-102. doi: 10.22071/gsj.2022.296260.1922.
Tabatabae, S.M. et al (2022). Prediction of the peak ground acceleration for Zagros earthquakes using ANFIS and data partitioning approach. Journal of Geography and Environmental Studies. 42 (11), 92-104.
Viganò, D.; Gabriele, S. & Dell'Arciprete, P. (2021). Earthquake risk assessment using augmented data. Earthquake Engineering and Structural Dynamics. 50 (7), 1805-1820.
Yin, Z.; Zhang, Y. & Zhang, J. (2020). Machine learning in the assessment of natural disaster risks: A review. Environmental Monitoring and Assessment. 192 (12), 786.
Wang, Z. et al (2020). MSLWENet: A novel deep learning network for lake water body extraction of google remote sensing images. Remote Sensing. 12, 4140.
Zhan, Q.; Miao, L. & Xiao, J. (2021). Intelligent algorithms for natural disaster prediction and management: Current status and future prospects. Journal of Natural Hazards. 105 (1), 123-142.
Zhang, X.; Xu, Y. & Chen, Y. (2020). Deep learning for flood prediction: A review. Natural Hazards. 105 (2), 123-146.