IAES International Journal of Artificial Intelligence (IJ-AI)

Vol. 12, No. 3, September 2023, pp. 1026~1032

ISSN: 2252-8938, DOI: 10.11591/ijai.v12.i3.pp1026-1032  1026

Earthquake prediction technique: a comparative study

Nada Badr Jarah1, Abbas H. Hassin Alasadi2, Kadhim Mahdi Hashim3

1
Statistics Department, General Specialty: Computer Science, College of Management and Economics, University of Basrah,
Basrah, Iraq
2
Information Systems Department, College of Computer Sciences and Information Technology, University of Basrah, Basrah, Iraq
3
Computer Technology Engineering Department, College of Information Technology, Imam Ja'afar Al-Sadiq University, Baghdad, Iraq

Article Info ABSTRACT

Article history: Earthquakes are one of the most dangerous natural disasters facing humans
because of their occurrence without warning and their impact on their lives
Received Aug 14, 2022 and property. In addition, predicting seismic movement is one of the main
Revised Oct 13, 2022 research topics in seismic disaster prevention. In geological studies, scientists
Accepted Dec 21, 2022 can predict and know the locations of earthquakes in the long term. Therefore,
about 80% of the major global earthquakes lie along the Pacific Ring belt,
known as the Ring of Fire. Machine learning methods have also been used for
Keywords: short-term earthquake prediction, and studies have applied the random forest
method to determine the factors that precede earthquakes. The machine
Earthquakes learning method was based on various decision trees, each of which predicted
Geology science the time to the nearest oscillation. The third group of scientists used the hybrid
Hybrid prediction method, which combines machine learning and geological studies.
Machine learning This research deals with a review of most of the geological studies and
Prediction machine learning techniques applied to earthquake data sets, which showed a
total lack of prediction of potential earthquakes through one approach, so
studies designed by geologists were combined with machine learning.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Nada Badr Jarah
Department of Statistics, Faculty of Management and Economics, University of Basrah
General Specialty: Computer Science
Basrah, Iraq
Email: [email protected]

1. INTRODUCTION
Earthquake is an important natural phenomenon affecting an organism's life and property. It is the
sudden release of energy transmitted by waves from the ground, it destroys large areas in a few minutes and
leads to huge losses in lives and property, and the idea of predicting earthquakes gives at least a little time to
protect people and reduce earthquake damage [1]. The seismic movement forecast innovation is being created
to foresee the vibration caused by seismic tremors to contribute to progressing the exactness of the seismic
movement forecast innovation utilised in seismic tremor chance appraisal to prepare for future seismic tremor
catastrophes and seismic tremor early caution promptly after the seismic tremor. This will lead to further
improvement, prevention and mitigation of earthquake disasters, represented in the continuous comparison
between geological studies and machine learning (ML) [2]. Geologists consider earthquakes a difficult task,
and their probable prediction in a given period is based on knowledge of all the data on the tectonic activity of
a region. By recording the Earth's seismic movement, seismologists try to obtain information about the physical
processes inside the Earth. The central target of attention has historically been the source of the earthquake [3].
Scientists applied a method of earthquake prediction using data recorded from a number of seismic stations
and machine learning methods to determine the factors that precede an earthquake based on the random forest

Journal homepage: http://ijai.iaescore.com

Int J Artif Intell ISSN: 2252-8938  1027

method. It was based on different decision trees, and machine learning algorithms were used to train and
analyse a pattern of acoustic data to predict earthquake occurrence [4].
Powerful computational techniques for big data analysis have emerged, and scientists have also been
able to apply a hybrid combination of machine learning and seismic prediction formulas based on traditional
physical models to improve the accuracy of seismic motion prediction while solving the problem of unbalanced
data learning that has a significant bias in such data [5]. This study was represented by a survey of previous
studies of earthquake geologists in observing the tectonic activity of the plates in the region, the seismic history
and various other studies. Nevertheless, accurate prediction of the timing of earthquakes has been difficult to
achieve, so researchers in the field of machine learning worked after recording readings of earthquake data by
designing and developing algorithms and techniques that allow computers to have the learning feature in
analysing sound and seismic waves of plate movement and predicting an earthquake with the simulation
process. Both the geological and informational studies did not reach the accuracy of predicting short-term
earthquakes unless the two studies were combined in one approach, and in terms of this research, a review of
earthquake prediction studies for each of the two studies and the combination of the two studies in a hybrid
study to improve the final performance of earthquake prediction to a large extent. To this end, the review
included the studies within three tables for each approach.

2. RELATED WORK
Researchers are doing their best to predict earthquakes, and several studies have presented different
forecasting methods and compared them to find the best prediction results. The following are the most
important studies related to the topic of research in comparing the most successful methods,the study by
Maqsoom et al. (2022) suggested two integrated frameworks: analytic network process (ANP)- artificial neural
network (ANN) and ANP- convolutional neural network (CNN), and 16 factors contributing to earthquake risk
were selected. Using geographic information system (GIS) to formulate it, a database was created for training
and testing models, thus designing earthquake hazards in North Pakistan, and the area under the curve (AUC)
values for ANN and CNN were 0.843 and 0.878, respectively, and this shows good performance [6]. The study
aims for Tehseen et al. (2020) to identify and compare the methods, models, frameworks and tools used to
predict earthquakes using criteria based on 70 studies published in 2010-2020. It showed that most of the
proposed models were long-term predictions. An analysis was conducted based on bibliometric and meta-
information by classifying articles according to research type, experimental type, approach, target area, and
system-specific parameters [3]. The study by Ogata (2013) describes the prospects for research in the ability to
predict earthquakes to achieve scientific prediction soon, techniques for predicting earthquakes based on
anomalies have been proposed, and we find that there is growing momentum for seismologists to develop an
organised research program on the possibility of exploring possibilities in earthquake prediction [7]. In the
study developed by Pushan et al. (2012), The method of linking the evaluative parameters to the analysis of
the data set used and the success rate of 18 of the most applicable algorithms was developed by comparing
different models for earthquake prediction [8].

3. METHOD
Earthquakes are part of the life of the Earth and a sad part of human history, and they are sudden
disturbances in the Earth's crust, as not a year goes by without hearing dozens of earthquakes, some of them
strong and destructive, wiping out entire cities, so the question of predicting their occurrence seems very
important. Predicting an earthquake means making an accurate prediction based on three factors: when the
earthquake will occur, where it will occur, and how large it will be. Until recently, one of the challenges
researchers faced was how to predict earthquakes and deal with a natural phenomenon. In the study [9] that
there is no valid prediction in the short term, as the reason for the short-term expectation is to enable crisis
measures to reduce traffic and destruction, leading to false prediction and disappointment to give caution in
the event of a major earthquake tremor may lead to legal obligation or habit vice versa bad in case of a false
warning of seismic tremor [10]. It is necessary to improve the spatial and temporal resolution and accuracy of
seismic activity prediction algorithms based on statistical and physical models. Furthermore, build a
mechanism for evaluating and validating those algorithms. In order to efficiently implement these matters, it
is essential to continue organically combining software, seismic data quality control, and seismic activity
prediction algorithms. Statistical hypothesis testing methods and ML approaches, namely, polynomial logistic
regression and the support vector machine (SVM) for earthquake data, may be used in regression classification
and analysis to determine the probability of an earthquake [11].

Earthquake prediction technique: a comparative study (Nada Badr Jarah)

1028  ISSN: 2252-8938

4. PREDICTION OF EARTHQUAKES USING MACHINE LEARNING

The world started using ML at the beginning of this century to predict earthquakes, where ML makes
accurate data-based predictions of events and factors affecting each other in a complex way. However, it is
indicated that ML is not good at predicting events that are not included in the training data often and occur
infrequently. If there is a large bias in the training data, then the prediction through ML will also be biased [12].
The application of unsupervised ML to analyse the full expression of earthquakes in these catalogs is the fastest
way to improve earthquake prediction; recent years have seen accelerating efforts in applying ML to earthquake
problems due to large data sets that will constitute a new generation of earthquake catalogs as well as
computational power [13]. The amount of data should be sufficient when analysing and predicting information
about earthquakes using any technique of ML. There is a problem that it is not possible to secure enough data,
especially for learning deep learning models, because large earthquakes are rare events. Compared with the
frequency of events observed in normal times, the low frequency of disasters is inevitable, so it is necessary to
take some measures in practice. It is worth noting that earthquakes of (7-7.9) on the Richter scale, which cause
great damage over a wide area, occur in the world every 18 years and that earthquakes of (8-9.9) on the Richter
scale, which cause great damage up to thousands of miles, occur once one in the world every 20 years [14].
Table 1 presents earthquake prediction studies using ML methods.

Table 1. Machine learning studies for earthquake prediction

Algorithm The method used Results
ANN − Data acquisition, preprocessing, feature − Studying the data collected from previous
extraction, and neural network training and earthquakes provides a better prediction accuracy
testing [9]. of 32%.
− Seismic coefficients are used, the b value, − The high success rate achieved supports the
Bath's law, and Omori - Utsu's law [15]. suitability of the soft computing application.
Support vector − By developing this mathematical model − The accuracy obtained for the stacking model is
regressor (SVR) and based on the training data set [16]. higher at 83%.
random forest
Supervised machine- − Using earthquake catalogs for analysis and − The fastest way to improve earthquake prediction.
learning prediction [13].
Deep learning − Training a deep learning model with large − Transfer learning and meta-learning were
amounts of data using a CNN [17]. introduced to build general earthquake detection
− Use of historical seismic events SVM and models
CNN [18]. − The proposed strategy performs well without
physical planning, including vectors, as within the
conventional neural organise strategy.
Artificial bee colony − Using the ABC algorithm that simulates the − The experimental result showed that multilayer
(ABC) algorithm intelligent search behavior of a honeybee perceptron (MLP)-ABC performed better than
swarm [19]. MLP- backpropagation (BP) for time series data.

5. GEOLOGICAL EARTHQUAKE PREDICTION

An earthquake is an effect of destroying the fault in the epicenter. Predicting earthquakes is in answer
to the following questions: the first, where the earthquake will occur, and the second: what is its expected
intensity, by studying the historical and instantaneous seismic activity of the study area and drawing its seismic
maps. As for the answer to the third question, it is the most difficult, and no methods have been found so far
that enable accurate prediction of the time of earthquakes in a specific area, but some indicators are used as
new indications for earthquakes to happen soon. The change in the geomagnetic field [20], temperature [21],
groundwater level [22], radon emission [23], and unusual animal behaviors [24] included the emergence of
new cracks and the growth of previous cracks.
Various attempts have been made for earthquake prediction, but a very reliable prediction has not
yet been achieved. The root cause of significant uncertainties in predicting current earthquakes is the severe
lack of experience with previous earthquake events and precursor phenomena. However, the most important
reason is that we have to make predictions based on empirical methods due to insufficient understanding and
modeling of the earthquake generation process and irregular and multi-scale phenomena. If seismology
develops and it becomes possible to understand better the seismic phenomenon, its internal structure and the
movement of the Earth where earthquakes occur, earthquake prediction with a completely different approach
can be descriptive. Prediction efforts may be possible, and although there is a limit to the empirical approach, it
is possible to obtain sufficiently useful information by accumulating much information over a long period.
The seismic motion prediction equation predicted the seismic motion strength index [25], [26]. The
seismic motion prediction equation models the effects related to earthquake shaking based on knowledge of
geophysics, an equation obtained by performing regression analysis using previous recordings, assuming

Int J Artif Intell, Vol. 12, No. 3, September 2023: 1026-1032

Int J Artif Intell ISSN: 2252-8938  1029

functional form connects parameters such as distance. Because of the ease of computation, the earthquake
motion prediction equation is used for simple calculation of earthquake motion in the engineering field,
especially for earthquake risk assessment that requires a great deal of computation and for early warning of
earthquakes immediately after an earthquake speed of computation. Since the seismic motion prediction
formula expresses a physical model based on academic knowledge to date, it is believed that it will exhibit
some prediction performance even when predicting irregularly occurring events. In Table 2, earthquake
prediction studies using geological methods are presented.

Table 2. Geological studies for predict earthquakes

Algorithm The method used Results
Bayesian ML Create a probabilistic model to predict long- Research's ML models predict a new
range earthquakes for each seismic zone [27]. period of strong earthquakes over
several years for seismic zones.
Teaching machine learning in Re-measure disappointment times as a Google's ML competition platform,
geophysics portion of the seismic cycle, compare the Kaggle, was used to engage the
dissemination of preparing inputs, and test worldwide community of ML in all the
information. In expansion to giving logical world and ML applications in Earth
bits of knowledge into blame forms within the sciences.
research facility and their relationship to the
advancement of factual properties of related
seismic information [28].
GIS and the program (Geo tool) A geological and topographical study of the The data found that it was a natural
northern border region with the Islamic earthquake not caused by any human
Republic of Iran; via waveform analysis [29]. action.
Ground movement expectation The study is to support the modernisation of the The greatest danger is located in the
conditions for utilisation in building code and to meet the need in Iraq for a northern cities of Sulaymaniyah, Erbil,
probabilistic seismic danger contemporary assessment of seismic risks in Badra and Al-Amarah fault areas.
evaluations (PSHA) terms of spectral acceleration. [30].
Use of three seismic catalogs, namely The seismic history of Diyala Governorate The Badra-Amarah fault extends from
European Mediterranean Seismological during 10 years (2004-2014) [31]. the city of Al-Amarah in the southeast
Centre (EMSC), Incorporated Research to the city of Mandali in the northwest.
Institutions for Seismology (IRIS), and It is the most seismically active fault in
the Iranian Seismological Center Iraq.
(IRSC)
Seismic source parameters from Study of the seismic history and focal solutions The fault is about 200 km long and may
different catalogs and official stress to the Badra Al-Amarah fault. [32]. be as deep as the basement rocks,
inversion technique which are about 10 km long. The fault
extends to the northwest to reach the
city of Mandalay.
Ground motion prediction equation Consider comparing the proposed The cruel and single standard
(GMPE) demonstration with NGA-West2 models and deviations of the ground speeding up
examining the territorial variety in soil top are displayed, and the pseudo-
movement in terms of unearthly shape, size dispersed range speeding up reaction
scale, separate scale, profundity scale, arranges 5% for the cruel of the
blaming mode, and area impacts [33]. direction- independent even
component of ground movement
RotD50 for a ghostly period of 0.01–
10 s.
Polynomial logistic regression, This work uses historical data on medium to In predicting future earthquakes, the
SVM, and Naïve Bayes long-term earthquakes and compares the SVM excels and produces large
performance of machine learning algorithms distances and magnitudes for the
[34]. current earthquake report.

6. HYBRID METHOD
Forming the joint research of machine learning experts and seismologists, we succeeded in predicting
earthquakes, as ML detected small signals treated as noise in traditional seismology. However, many
difficulties must be overcome before they can be applied to earthquakes. The hybrid method combines the
advantages of a seismic survey and motion prediction formula based on the physical model, such as stability
in predicting rare events. It allows for more accurate predictions than a newly developed method [35].
It utilises both ML and geological studies, which enables highly flexible and accurate prediction
according to the data, and the seismic motion prediction formula, which ensures the prediction performance of
events that occur somewhat erratically based on physical models. To foresee seismic movement, we considered
a half-breed expectation strategy that combines all of them. Particularly, after the first-stage expectation is
performed, utilising the current seismic movement forecast equation; the second-stage forecast is actualised by
ML in a frame that completes the imperfect portion, the combination of which is the ultimate. Yield as an
expectation. Various methods have been used, for example, random trees as a ML method, one of the random

Earthquake prediction technique: a comparative study (Nada Badr Jarah)

1030  ISSN: 2252-8938

forest derivation algorithms that do collective learning using multiple decision trees that do classification and
regression using a tree structure, random forests perform decision tree branching during training to maximise
gains. In contrast, highly random trees do so at random. Table 3 shows a presentation of earthquake prediction
studies using hybrid methods.

Table 3. Studies of hybrid methods for earthquake prediction

Algorithm The method used Results
Regression algorithms combined Prediction of earthquake magnitude within Very promising results that help in
with group learning in the context the next seven days. The Apache Spark processing 400 huge data simultaneously
of big data framework, the water library in the R with a large number of variables.
language, and the Amazon cloud
infrastructure were used [36].
Deep learning model for It uses seismic indices (explicit features) Experimental results on eight datasets
earthquake prediction (DLEP) designed by geologists and feature vectors from different regions show the
combines explicit and implicit (implicit features) extracted from deep effectiveness of the proposed DLEP for
features learning methods [37]. earthquake prediction.
Highlight extraction (mRMR) and Sixty seismic highlights were calculated The prediction performance of all studied
crossover neural organize (HNN) utilising seismological concepts, such as regions improved compared to previous
(SVR) Gutenberg-Richter law, seismic rate varieties, prediction studies.
stun recurrence, seismic vitality discharge,
and add up to redundancy time [5].
Classification and regression tree Generate current mathematical and statistical Experimental results of two historical
(CART) algorithm classification features directly as seismic indicators. seismic records in China show the
and regression tree Moreover, a regression tree algorithm effectiveness of features based on initial
predicts the main shock's naming [38]. patterns proposed with the specific
CART algorithm for earthquake
prediction.
A hybrid approach to ML and the This hybrid approach of ML technology and Better prediction than any of the
traditional ground motion the physical model underestimates the individual methods applied alone
prediction equation importance of strong movements [39].

7. CONCLUSIONS
Geological studies in predicting earthquakes predict for a long period that may exceed months or
years, and it is approximate and inaccurate in determining the location, size, and time of the earthquake, as the
time of the fault movement is measured before and predicts the future in the frequency of earthquakes. The
result of using ML is to improve the performance of seismic motion prediction technology, and we believe that
ML will continue to grow and become more important in all areas of Earth sciences. Improved earthquake
prediction through historical seismic data has also been used. The most promising approach is using artificial
intelligence and ML to gain more knowledge, and just applying ML can cause problems when anticipating
events that occur infrequently. The mere application of ML may cause problems when predicting events that
occur irregularly. We found new horizons for seismic research in training ML algorithms with signs of an
upcoming earthquake based on the sound produced. This research approach to solve this problem by combining
the geophysical model and machine learning can be applied in other fields. In addition, earthquake prediction
requires dynamism in application and self-adaptation in adjusting to the inferred variables.

REFERENCES
[1] A. Aanuoluwa and S. Lukman Ayobami, “Earthquake: a terrifying of all natural Phenomena,” Journal of Advances in Biological
and Basic Research, vol. 1, no. June, pp. 4–11, 2015.
[2] R. M. Allen and D. Melgar, “Earthquake early warning: advances, scientific challenges, and societal needs,” Annual Review of
Earth and Planetary Sciences, vol. 47, pp. 361–388, 2019, doi: 10.1146/annurev-earth-053018-060457.
[3] R. Tehseen, M. S. Farooq, and A. Abid, “Earthquake prediction using expert systems: a systematic mapping study,” Sustainability
(Switzerland), vol. 12, no. 6, 2020, doi: 10.3390/su12062420.
[4] P. Tosi, P. Sbarra, and V. De Rubeis, “Earthquake sound perception,” Geophysical Research Letters, vol. 39, no. 24, 2012,
doi: 10.1029/2012GL054382.
[5] K. M. Asim, A. Idris, T. Iqbal, and F. Martínez-Álvarez, “Earthquake prediction model using support vector regressor and hybrid
neural networks,” PLoS ONE, vol. 13, no. 7, 2018, doi: 10.1371/journal.pone.0199004.
[6] A. Maqsoom, B. Aslam, U. Khalil, M. A. Mehmood, H. Ashraf, and A. Siddique, “An integrated approach based earthquake risk
assessment of a seismically active and rapidly urbanizing area in Northern Pakistan,” Geocarto International, 2022,
doi: 10.1080/10106049.2022.2105404.
[7] Y. Ogata, “A prospect of earthquake prediction research,” Statistical Science, vol. 28, no. 4, pp. 521–541, 2013,
doi: 10.1214/13-STS439.
[8] M. N. K. Pushan, O. Mishra, “Decision analysis for earthquake prediction methodologies: fuzzy inference algorithm for trust
validation,” International Journal of Computer Applications, vol. 45, no. 4, pp. 13–20, 2012, doi: 10.5120/6767-9048.
[9] A. S. N. Alarifi, N. S. N. Alarifi, and S. Al-Humidan, “Earthquakes magnitude predication using artificial neural network in northern

Int J Artif Intell, Vol. 12, No. 3, September 2023: 1026-1032

Int J Artif Intell ISSN: 2252-8938  1031

Red Sea area,” Journal of King Saud University - Science, vol. 24, no. 4, pp. 301–313, 2012, doi: 10.1016/j.jksus.2011.05.002.
[10] Y. Y. Kagan, “Are earthquakes predictable?,” Geophysical Journal International, vol. 131, no. 3, pp. 505–525, 1997,
doi: 10.1111/j.1365-246x.1997.tb06595.x.
[11] K. S. Riedel, “Statistical tests for evaluating earthquake prediction methods,” Geophysical Research Letters, vol. 23, no. 11,
pp. 1407–1409, 1996, doi: 10.1029/96GL00476.
[12] M. H. Al Banna et al., “Application of artificial intelligence in predicting earthquakes: state-of-the-art and future challenges,” IEEE
Access, vol. 8, pp. 192880–192923, 2020, doi: 10.1109/ACCESS.2020.3029859.
[13] G. C. Beroza, M. Segou, and S. Mostafa Mousavi, “Machine learning and earthquake forecasting—next steps,” Nature
Communications, vol. 12, no. 1, 2021, doi: 10.1038/s41467-021-24952-6.
[14] R. Jena, B. Pradhan, G. Beydoun, A. Al-Amri, and H. Sofyan, “Seismic hazard and risk assessment: a review of
state-of-the-art traditional and GIS models,” Arabian Journal of Geosciences, vol. 13, no. 2, p. 50, Jan. 2020,
doi: 10.1007/s12517-019-5012-x.
[15] J. Reyes, A. Morales-Esteban, and F. Martínez-Álvarez, “Neural networks to predict earthquakes in Chile,” Applied Soft Computing
Journal, vol. 13, no. 2, pp. 1314–1328, 2013, doi: 10.1016/j.asoc.2012.10.014.
[16] P. Bangar, D. Gupta, S. Gaikwad, B. Marekar, and J. Patil, “Earthquake prediction using machine learning algorithm,”
International Journal of Recent Technology and Engineering (IJRTE), vol. 8, no. 6, pp. 4684–4688, Mar. 2020,
doi: 10.35940/ijrte.E9110.018620.
[17] J. Audretsch, “Earthquake detection using deep learning based approaches,” King Abdullah University of Science and Technology
Thuwal, 2020, [Online]. Available: https://www.ptonline.com/articles/how-to-get-better-mfi-results.
[18] E. Amfo, “Earthquake magnitude prediction using support vector machine and convolutional neural network,” Open Access Theses
& Dissertations, 2019, [Online]. Available: https://digitalcommons.utep.edu/open_etd/1970.
[19] R. G. H. Shah and N. Nawi, “Using artificial bee colony algorithm for MLP training on earthquake time series data prediction,”
Journal of Computing, vol. 3, no. 6, pp. 135–142, 2011.
[20] S. K. Park, “Precursors to earthquakes: seismoelectromagnetic signals,” Surveys in Geophysics, vol. 17, no. 4, pp. 493–516, 1996,
doi: 10.1007/BF01901642.
[21] S. C. Mavrodiev et al., “Study of the possibility of predicting earthquakes,” International Journal of Geosciences, vol. 09, no. 12,
pp. 688–706, 2018, doi: 10.4236/ijg.2018.912042.
[22] M. Senthilkumar, D. Gnanasundar, B. Mohapatra, A. K. Jain, A. Nagar, and P. K. Parchure, “Earthquake prediction from high
frequency groundwater level data: a case study from Gujarat, India,” HydroResearch, vol. 3, pp. 118–123, 2020,
doi: 10.1016/j.hydres.2020.10.004.
[23] H. A. Khan, M. Tufail, and A. A. Qureshi, “Radon signals for earthquake prediction and geological prospection,” Journal of Islamic
Academy of Sciences, vol. 3, no. 3, pp. 229–231, 1990.
[24] V. K. Katiyar, M. Sharma, N. Bhargava, M. L. Sharma, and P. Pradhan, “Earthquake prediction through animal behavior: a review,”
Indian Journal of Biomechanics, vol. 78, no. Special Issue, pp. 159–165, 2009.
[25] J. J. Bommer, J. Douglas, F. Scherbaum, F. Cotton, H. Bungum, and D. Fäh, “On the selection of ground-motion prediction
equations for seismic hazard analysis,” Seismological Research Letters, vol. 81, no. 5, pp. 783–793, 2010,
doi: 10.1785/gssrl.81.5.783.
[26] B. Gutenberg and C. F. Richter, “Earthquake magnitude, intensity, energy, and acceleration,” Bulletin of the Seismological Society
of America, vol. 32, no. 3, pp. 163–191, 1942, doi: 10.1785/bssa0320030163.
[27] V. M. Velasco Herrera et al., “Long-term forecasting of strong earthquakes in North America, South America, Japan,
Southern China and Northern India with machine learning,” Frontiers in Earth Science, vol. 10, 2022,
doi: 10.3389/feart.2022.905792.
[28] P. A. Johnson et al., “Laboratory earthquake forecasting: a machine learning competition,” Proceedings of the National Academy
of Sciences of the United States of America, vol. 118, no. 5, 2021, doi: 10.1073/pnas.2011362118.
[29] A. Saeel, “Using modern technology in analyzing earthquake on the Iraq-Iran Borderline: South-East Halabja-2017 2018,”
AL- AMEED JOURNAL, vol. 10, no. 1, pp. 355–386, 2021.
[30] W. Abdulnaby et al., “Probabilistic seismic hazard assessment for Iraq,” Journal of Seismology, vol. 24, no. 3, pp. 595–611, 2020,
doi: 10.1007/s10950-020-09919-2.
[31] W. Abdulnaby, R. Al-Mohmed, and M. Mahdi, “Seismicity and recent stress regime of Diyala City, Iraq–Iran border,” Modeling
Earth Systems and Environment, vol. 2, no. 3, 2016, doi: 10.1007/s40808-016-0201-z.
[32] R. A.-M. W. Abdulnaby, M. Mahdi and H. Mahdi, “Seismotectonic of Badra-Amarah fault, Iraq-Iran border,” IOSR Journal of
Applied Geology and Geophysics (IOSR‐JAGG), vol. 4, no. 3, pp. 27–33, 2016.
[33] V. B. Phung, C. H. Loh, S. H. Chao, B. S. J. Chiou, and B. S. Huang, “Ground motion prediction equation for crustal earthquakes
in Taiwan,” Earthquake Spectra, vol. 36, no. 4, pp. 2129–2164, 2020, doi: 10.1177/8755293020919415.
[34] I. M. Murwantara, P. Yugopuspito, and R. Hermawan, “Comparison of machine learning performance for earthquake prediction in
Indonesia using 30 years historical data,” Telkomnika (Telecommunication Computing Electronics and Control), vol. 18, no. 3,
pp. 1331–1342, 2020, doi: 10.12928/TELKOMNIKA.v18i3.14756.
[35] V. G. Kossobokov, “Earthquake prediction: basics achievements, perspectives,” Acta Geodaetica et Geophysica Hungarica,
vol. 39, no. 2-3 SPEC. ISS., pp. 205–221, 2004, doi: 10.1556/AGeod.39.2004.2-3.6.
[36] G. Asencio-Cortés, A. Morales-Esteban, X. Shang, and F. Martínez-Álvarez, “Earthquake prediction in California using regression
algorithms and cloud-based big data infrastructure,” Computers and Geosciences, vol. 115, pp. 198–210, 2018,
doi: 10.1016/j.cageo.2017.10.011.
[37] R. Li, X. Lu, S. Li, H. Yang, J. Qiu, and L. Zhang, “DLEP: a deep learning model for earthquake prediction,” Proceedings of the
International Joint Conference on Neural Networks, 2020, doi: 10.1109/IJCNN48605.2020.9207621.
[38] L. Zhang, L. Si, H. Yang, Y. Hu, and J. Qiu, “Precursory pattern based feature extraction techniques for earthquake prediction,”
IEEE Access, vol. 7, pp. 30991–31001, 2019, doi: 10.1109/ACCESS.2019.2902224.
[39] H. Kubo, T. Kunugi, W. Suzuki, S. Suzuki, and S. Aoi, “Hybrid predictor for ground-motion intensity with machine
learning and conventional ground motion prediction equation,” Scientific Reports, vol. 10, no. 1, 2020,
doi: 10.1038/s41598-020-68630-x.

Earthquake prediction technique: a comparative study (Nada Badr Jarah)

1032  ISSN: 2252-8938

BIOGRAPHIES OF AUTHORS

Prof. Nada Badr was born in Basra, Iraq, on March 24, 1961. She received her
MSc in Computer Science from Basra University in 2002, and she is currently a professor at
the College of Administration and Economics - Department of Statistics- University of
Basrah, Basrah. She can be contacted at email: [email protected].

Abbas H. Hassin Alasadi is a Professor at the Computer Information Systems

Department, College of Computer Science and Information Technology, University of Basrah,
Basrah, Iraq. He received his PhD from the School of Engineering and Computer Science /
Harbin Institute of Technology, China. He spent more than ten years as Associate Professor
at different Universities abroad, his current position. His research interests include medical
image processing, biometrics, information retrieval, and human-computer interaction. His
research work has been published in various international journals and conferences. Abbas is
an active reviewer in many computer sciences and software engineering journals. He is one of
the ACIT, UJCS, SIVP, and IJPRAI reviewer members. He can be contacted at email:
[email protected].

Prof. Dr Kadhim Mahdi was born in the city of Dhi Qar, Iraq, on November
25, 1962. He obtained a doctorate in computer sciences from Basrah University's Computer
Science Department at the College of Science in 2006. In 2012, he was promoted to Professor.
In the areas of image processing, information security, pattern recognition, and network
security. He is currently a professor at Imam Jaafar Al-Sadiq University in Baghdad. He can
be contacted at email: [email protected].

Int J Artif Intell, Vol. 12, No. 3, September 2023: 1026-1032