Elite GA-based Feature Selection of LSTM for

Earthquake Prediction
Zhiwei Ye
Hubei University of Technology
Wuyang Lan
Hubei University of Technology
Wen Zhou (  [email protected] )
Hubei University of Technology
Qiyi He
Hubei University of Technology
Liang Hong
Hubei Provincial Geographical National Conditions Monitoring Center
Xinguo Xu
Hubei Provincial Geographical National Conditions Monitoring Center
Yunxuan Gao
Hubei University of Technology

Research Article

Keywords: Earthquake magnitude prediction, Elite genetic algorithm, Long short-term memory, AETA

Posted Date: June 19th, 2023

DOI: https://doi.org/10.21203/rs.3.rs-3049982/v1

License:   This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License

Additional Declarations: No competing interests reported.

 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM

for Earthquake Prediction
Zhiwei Ye1 , Wuyang Lan1 , Wen Zhou1*, Qiyi He1 , Liang
Hong2 , Xinguo Xu2 and Yunxuan Gao1
1* School of Computer Science, Hubei University of Technology,
No. 28, Nanli Road, Hongshan District, Wuhan, 430068, Hubei,
China.
2 Hubei Provincial Geographical National Conditions Monitoring

Center, No. 50, Zhongnan 1st Road, Wuhan, 430064, Hubei,

China.

*Corresponding author(s). E-mail(s): zw [email protected];

Contributing authors: [email protected]; [email protected];
[email protected]; [email protected]; [email protected];
[email protected];

Abstract
Earthquake magnitude prediction is an extremely difficult task that
has been studied by various machine learning researchers. However, the
redundant features and time series properties hinder the development
of prediction models. Elite Genetic Algorithm (EGA) has the advan-
tages in searching optimal feature subsets, meanwhile, Long Short-Term
Memory (LSTM) is dedicated to processing time series and complex
data. Therefore, we propose an EGA-based feature selection of LSTM
model (EGA-LSTM) for time series earthquake prediction. First, the
acoustic and electromagnetics data of the AETA system we developed
are fused and preprocessed by EGA, aiming to find strong correlation
indicators. Second, LSTM is introduced to execute magnitude pre-
diction with the selected features. Specifically, the RMSE of LSTM
and the ratio of selected features are chosen as fitness components of
EGA. Finally, we test the proposed EGA-LSTM on the AETA data
of Sichuan province, including the influence of data in different peri-
ods (timeP eriod) and fitness function weights (ωa and ωF ) on the
prediction results. Linear Regression (LR), Support Vector Regression
(SVR), Adaboost, Random Forest (RF), standard GA (SGA), steadyGA,

1
 Springer Nature 2021 LATEX template

2 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

and three Differential Evolution Algorithms (DEs) are adopted as our

baselines. Experimental results demonstrate that all the methods can
get the best performance when timeP eriod = 0 : 00 − 8 : 00,
ωa = 1, and ωF = 0.8. Moreover, our proposed approach is supe-
rior to state-of-the-art approaches on the evaluation indicators MAE,
MSE, RMSE, and R2 . Non-parametric tests reveal that EGA-LSTM is
significantly different from others and outperforms the standard LSTM.

Keywords: Earthquake magnitude prediction, Elite genetic algorithm, Long

short-term memory, AETA

1 Introduction
Earthquakes are serious, sudden-onset natural disasters. Since 1970, at least 3.3
billion people have died from such natural disasters, and 226 million people are
directly affected each year [1]. Earthquakes cause serious human injuries and
economic losses due to the destruction of buildings and other rigid structures.
Nevertheless, earthquake prediction is essential and necessary to reduce human
casualties and economic loss.
Earthquake Moment Magnitude (MW) prediction is one of the important
issues in earthquake prediction fields. Many studies have been carried out for
finding the relationships in earthquakes and MW prediction. However, the
complex and non-linear relationships in earthquakes make MW prediction to
be a challenging task. Machine Learning (ML) methods have been adopted for
MW prediction due to their high classification accuracy. Adeli et al. [2] cal-
culated the seismic historical indicators, including Gutenberg Richer b-values,
time lag, earthquake energy, and mean magnitude based on ML methods to
predict MW. Asim et al. [3] investigated the Cyprus earthquake catalog tem-
porally and computed sixty seismic features. Then, these features served as
the input instances of Support Vector Machines (SVM) and RF to predict five
days-ahead, one week-ahead, ten days-ahead, and fifteen days-ahead with dif-
ferent MW thresholds. Crustal movement is a continuous process makes MW
prediction to be a time-series issue. Hence, in [4], Berhich et al. calculated an
appropriate feature by historical data to enhance time-series task and imple-
mented LSTM to predict MW with the enhanced features. Cai et al. [5] used
three groups of real seismic data: gravity, georesistivity, and water-level dataset
to predict MW based on LSTM. They considered the time series of precur-
sors and earthquakes, however, there are irrelevant or even redundant features
within the precursors, resulting in low prediction accuracy.
Feature Selection (FS) is a critical issue to improve the classification
accuracy of ML methods. FS can be regarded as combinatorial optimization
problems and EGA has the advantage in searching optimal feature subset.
Kadam et al. [6] proposed EGA for selecting features in arrhythmia classifica-
tion, and achieved satisfied performance. Thus, we adopt EGA for FS in MW
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 3

prediction and design a novel fitness function, which is calculted by the RMSE
of LSTM and the ratio of selected features. Finally, considering the time-series
effect of electromagnetic and acoustic data from AETA and the abruptness of
earthquake, we introduce LSTM to predict magnitude with optimal feature
subset selected by EGA.
To verify the performance of the proposed EGA-LSTM, 95 features that
are derived from the electromagnetic and acoustic data collected by AETA are
adopted as our experimental dataset. After the feature selection process by
EGA, the chosen features served as the input data of LSTM model. Besides,
five Evolution Algorithms (EAs) and four different ML methods are adopted
as our baselines: SGA, steadyGA, and three DEs, LR, SVR, Adaboost, and
RF. Five ML evaluation indicators are adopted to assess the performance
of those methods. The experimental results demonstrate that our proposed
EGA-LSTM is superior to the state-of-the-art methods.
The main contributions of this study are as follows:
• The original dataset is collected by the AETA system developed by our
team, which detects electromagnetic and acoustic to predict earthquakes in
Sichuan and surroundings.
• To eliminate redundant features from the original 95 features, we propose
EGA with a novel fitness function of a specific seismic scene to find the
optimal solution.
• Considering earthquake magnitude’s time series effect and the abruptness
of earthquake, we choose LSTM as prediction model. Other ML methods,
such as LR, SVR, Adaboost, and RF are our baselines.
• The statistical measures (MAE, MSE, RMSE, R-square) are used to measure
the performance of suggested ML models.
The remainder of the study is arranged as follows. Section 2 depicts the
latest and related works. The method will be fully described in section 3. In
section 4, the experiments of EGA-LSTM and other EAs and ML methods
will be illustrated in detail. The conclusion and future works will be shown in
section 5.

2 Related Work
Various approaches have been implemented in earthquake prediction, including
feature selection methods and prediction methods.

2.1 Seismic feature selection

Feature selection is a crucial step in building a robust model, especially in high-
dimensional data scene [7]. The application of feature selection to earthquake
prediction is a complex problem, because different techniques are sensitive to
data imbalance and noise. Martı́nez-Álvarez et al. [8] proposed information
gain to evaluate different indicators and removed the low-ranked or null con-
tribution seismic indicators. Roiz-Pagador et al. [9] introduced four different
 Springer Nature 2021 LATEX template

4 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

feature selection methods and nine different EAs in correlation-based feature

selection methods.

2.2 Earthquake prediction methods

The main MW prediction methods contain mathematical models, shallow ML
techniques, and DL methods.
For mathematical models, the mathematical methods have been imple-
mented to deal with uncertainty problems, such as probability theory and
fuzzy sets to predict MW. For instance, Chen et al. [10] focused on the laws
of the earthquake time series based on chaos theory and carried on the earth-
quake forecast simulation through the analysis of real data. The quantitative
identification of time-series data result shows that seismic time-series data
perform deterministic chaotic characteristics. Cekim et al. [11] captured the
dynamics of earthquake occurrence using a novel Singular Spectrum Analysis
and Autoregressive Integrated Moving Average method to predict the aver-
age and maximum MW in the East Anatolian Fault of Turkey. However, the
prediction results proved unsatisfactory due to the strict requirements of the
mathematical model on the independence and correlation of the data.
Another branch of MW prediction approaches focused on shallow ML tech-
niques. They are driven by data and don’t require many parameters. Since
there is no definite mathematical relationship between precursor indicators
and location and MW, Panakkat et al. [12] introduced three different neural
networks: feed-forward Levenberg-Marquardt Back Propagation, RNN, and
Radial Basis Function (RBF) neural network to predict MW and location
with eight indicators from Gutenberg-Richter power-law. In 2017, Asim et
al. [13] proposed different ML approaches including RF and Linear Program-
ming Boost ensemble of decision trees for earthquake prediction in the region
of Hindukush. These techniques show significant and encouraging results but
they were still not available so far. As Chile is rocked by intraplate, inter-
face, and crustal events, Chanda et al. [14] proposed six different ML methods
to predict the total duration and significant duration for the three types of
earthquakes. Shah et al. [15] introduced an Improved Artificial Bee Colony
algorithm to improve the training process of Multilayer Perceptron. Muham-
mad et al. [16] studied earthquakes, Radon gas, and the Ionospheric total
electron content and used statistical and machine/deep learning methods to
find the relationship among the above-mentioned three elements in North
Anatolian Fault. Yang et al. [17] proposed an automated regression pipeline
approach for high-efficiency earthquake prediction. The core prediction part
is training RF with bayesian parameter optimization. Ensemble learning has
also been used in the field of earthquake prediction. Asim et al. [18] proposed
a novel earthquake predictor system combining seismic indicators with GA
and AdaBoost. GA’s searching capabilities and boosting of AdaBoost makes
the proposed method to be a powerful classifier. In addition, considering the
earthquake prediction process is similar to the anomaly detection process of
the biological immune system, Zhou et al. [19][20][21][22] introduced dendritic
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 5

cells algorithm, artificial macrophage classification optimization method, and

artificial antigen-presenting cells approach to earthquake prediction. However,
ML methods are unable to learn the complex and nonlinear relationship of
earthquake features. Therefore, we need a more powerful technique to solve
this problem.
Deep learning methods are experts at solving nonlinear problems because
they have many hidden layers and densely connected neurons that preserve
complex information. Moustra et al. [23] applied Artificial Neural Network
with the history of MW data and seismic electric signals in Greece to predict
MW, but the result is not satisfactory. In 2018, Asim et al. [24] computed
sixty features by employing seismological concepts, extracted features by Max-
imum Relevance and Minimum Redundancy (mRMR), and predicted MW
with the selected features based on SVR and HNN. In addition, HNN is sup-
ported by enhanced Particle Swarm Optimization. Jain et al. [25] proposed a
DL method to predict the position and depth of earthquakes. They analyzed
the effects of parameters on different datasets. However, these methods didn’t
catch the time series effects and only utilize DL methods to learn the rela-
tionship between features and magnitude. Draz et al. [26] believed that the
evolution and appearance of earthquake precursors exhibit complex behavior.
Therefore, they used deep machine learning for the detection of ionosphere and
atmosphere precursors. This study showcases the importance of machine learn-
ing techniques in earthquake detection, which contributes to the understanding
of the Lithosphere-Atmosphere-Ionosphere Coupling mechanism. Berhich et
al. [27] presented LSTM for earthquake prediction to study the correlations
between two divided groups considering their range of MW. To find the specfic
pattern among magnitude, timing and location. Berhich et al. [28] used LSTM
to learn temporal relationships, and the attention mechanism extracts impor-
tant patterns and information from input features. Kavianpour et al. [29]
introduced a novel prediction model based on the attention mechanism, Con-
volution Neural Network (CNN), and Bidirectional Long Short-Term Memory
(BiLSTM) models, to capture the temporal dependencies, which can predict
the maximum MW and the number of earthquakes in a specified period in
mainland China. Berhich et al. [4] calculated an appropriate feature to enhance
time feature and predict MW based on LSTM with this feature. Nevertheless,
most of these DL methods didn’t consider some precursor features are irrele-
vant features, even redundant features. These irrelevant features can make DL
models difficult to execute learning process, reducing predictive accuracy.
Generally speaking, the MW prediction process contains two steps: find-
ing a valid optimization precursor subset and constructing a predicting model.
Before a major earthquake stroke, it is often accompanied by changes in elec-
tromagnetic and acoustic signals in AETA. Hence, we have reasons to prove
that electromagnetic and acoustic signals are related to MW. As far as authors
know, electromagnetic and acoustic signals as precursor features haven’t been
 Springer Nature 2021 LATEX template

6 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

researched by LSTM, so we utilize these indicators to predict MW. In addi-

tion, considering the existence of redundant features, we propose EGA-LSTM
for time series earthquake prediction based on AETA.

3 Methodology
This section describes the proposed model to perform earthquake prediction.
EGA-LSTM is presented to predict MW.

3.1 Model overview

A schematic presentation of the proposed model is shown in Fig. 1. First,
the dataset from AETA was detected in the Sichuan and surroundingd. This
dataset includes two types, electromagnetic and acoustic, with a total of 95 fea-
tures. AETA has recorded electromagnetic and acoustic signals from January
1, 2017 to December 31, 2022 and detects them every ten minutes. So, there
are more than 20000 records. Since MW predicting in a short period makes
less sense and human activities affect electromagnetic and acoustic signals, we
merge signals from one day into one piece of catalog. The fusing detail will be
described in section 3.2. Then, we proposed EGA for feature selection. As Fig.
1 depicts, the selected features are divided into two groups: the training set and
the testing set. The time sequence of the train set and test set don’t change.
In this work, EGA-LSTM has been compared to other classic ML methods,
such as LR, SVR, Adaboost, and RF. Finally, different evaluation indicators
are applied to evaluate these algorithms.

Fig. 1 The program diagram of the proposed earthquake prediction model

3.2 Process of EGA-LSTM

The EGA-LSTM architecture for earthquake magnitude prediction depicts as
Fig. 2, meanwhile, the pseudocode of EGA-LSTM is shown in Algorithm 1.
Their main idea is as follows. The AETA dataset includes 95 features, 51
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 7

features for electromagnetic signals, and 44 features for acoustic signals. How-
ever, MW isn’t detected in AETA. Hence, the MW data is chosen from China
Earthquake Networks Center. As a result, the algorithm merges 95 features
and MW according to time. Since the time interval is short and MW prediction
in a short period makes less sense. Hence, we select a typical value in every
feature to represent a whole day. That means the size of the original one-day
data is 144*96. After fusing phase, the size becomes 1*96. In addition, electro-
magnetic and acoustic signals may be affected by human activities, we need
to choose a suitable timeP eriod to present a whole day. Considering the lack
of nighttime activities, we choose the data from the timeP eriod 0:00 to 8:00.
Then, the algorithm transforms time series data into supervised learning. In
this algorithm, the input variable is electromagnetic and acoustic signals and
the output variable is the MW of the following day. The step size is equal to
1. That means the algorithm can predict MW of the next day.

Fig. 2 The EGA-LSTM architecture for earthquake magnitude prediction

After processing the original data, the algorithm executes feature selection.
This model adopts EGA to select optimal feature subset, the individual of
EGA is applied to binary encoding and the dimension of each individual is
95. If one feature is selected, the corresponding dimension is 1, otherwise 0.
With comprehensive consideration of prediction accuracy and time complexity,
we proposed a novel fitness function defined as Eq. (1), where ωa and ωF is
 Springer Nature 2021 LATEX template

8 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

Algorithm 1 EGA-LSTM
Input: electromagnetic and acoustic signals from AETA
Output: MW of the next day
1: M axiter: The maximum number of iterations
2: Of s: The selected optimal feature subset
3: Use the max value of the data between 0:00 and 8:00 representing the
whole day
4: Time series transform into supervised learning sequence, step size equal to
1
5: Initialize population and parameter
6: while t < M axiter do
7: Calculate fitness for each individual
8: for each individual do
9: Retaining the best individuals
10: Selected operation
11: Crossover operation
12: Mutate operation
13: end for
14: Generate new population
15: end while
16: OF S = best f itness individual
17: M W = LST M (OF S)
18: return M W

the weight factor, F is the number of selected features, and P is the number
of all features. The first part of Eq. (1) represents the prediction accuracy
and the second part indicates the complexity of the model. In addition, Root
Mean Squared Error (RMSE) is an evaluation indicator that is defined as
Eq. (2), where ŷ is the predicting value and is the true value. This fitness
function enables considerable accuracy and the number of selected features is
significantly reduced.
F
f itness = ωa ∗ RM SE + ωF ∗ ( ) (1)
v P
u n
u1 X
RM SE(y, ŷ) = t k yi − ŷ k22 (2)
n i=1
The algorithm first defines the maximum number of iterations, crossover
rate, mutate rate, and the number of individuals. Then the model randomly
initializes the population and begins to iterate. The first step of iterating is
calculating the fitness of every individual. Hereafter, the algorithm executes
the selection operator, crossover operator, and mutation operator to generate
a new subpopulation. The loop finishes until the number of iterations reaches
the maximum number of iterations. After the iteration phase, the best indi-
vidual is the selected optimal feature subset. The last step of the algorithm
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 9

is predicting MW. At the beginning, the selected optimal feature subsets are
normalized by the Min-Max scaler, which is a transform method calculated
by Eq. (3). In this model, optimal feature subsets are mapped into the same
range which is between 0 and 1, which enables no one feature dominates the
others. The next step consists of dividing 70% of the training set and 30%
of the testing set. Then LSTM is trained and supported by RMSE for error
calculation and evaluation. Finally, after the model is trained, LSTM predicts
MW on the testing set, and the error between the real MW and the predicted
MW is calculated by different evaluation indicators. They are widely used to
evaluate regression models and applied to earthquake prediction. The detailed
evaluation indicators will be provided in section 4.2.

x − min(x)
z= (3)
max(x) − min(x)

3.3 Feature selection of EGA

GA is expert in solving combinatorial optimization problems since its operator
was designed based on discrete encoding [30]. Feature selection is a combina-
torial optimization problem and non-deterministic polynomial problem. The
traditional feature selection method is to find relevant features to the tar-
get variables. However, the features among the selected features subset should
be weakly correlated. Combining the above-mentioned two conditions, the
model can find the relationship between features and target variable. EGA
can directly take the optimal gene and directly participate in the next evolu-
tion without selection. Thus, we adopt EGA to find the optimal feature subset
based on GA’s searching ability.
In this method, we set the individual as a 95-dimensional vector and each
element is equal to 1 or 0. If the element is equal to 1, it means the cor-
responding dimension’s feature is selected, otherwise without being selected.
Initially, we initialize these individuals randomly. Then, calculate the fitness of
each individual based on Eq. (1). To get RMSE in Eq. (1), we train a sample
LSTM on the training test and predict MW on the testing set. Then calculate
RMSE based on Eq. (2). To decrease the time complexity and increase the
accuracy, we design a novel fitness function as Eq. (1). The first part repre-
sents the prediction accuracy and the second part represents time complexity.
After calculating the fitness, the algorithm begins to select individuals. Before
every loop, EGA retains the best individuals. The selection part is applied
to the roulette wheel to produce two new solutions. The algorithm gener-
ates a random probability value and selects the corresponding individual. The
next step is the crossover operator. The chosen crossover technique in this
model is double-point. There are two crossover points and the chromosomes
between two points are swapped only. The mutation operator is inversion
mutation randomly generating a gene locus in a low mutate rate and invert-
ing. If the old gene locus value is 1, the new value is 0. After generating a
new population, if the number of iterations reaches the maximum, output the
 Springer Nature 2021 LATEX template

10 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

best individual, otherwise recalculate fitness. The best individual gene (output
OF S = {of1 , of2 , ..., ofn }) is the optimal feature subset.

3.4 The LSTM for MW prediction

LSTM aims to solve the problem of long-term dependence [31]. Hence, we
apply LSTM to predict MW with the output OF S of EGA feature selection
phase. The diagram of LSTM for MW prediction is shown in Fig. 3. LSTM cell
is enhanced by three components called gates, including forget gate, update
gate, and output gate and two memory cells: hidden state and internal state.

Fig. 3 The LSTM for MW prediction

The selected features are processed by the Min-Max scaler and the squashed
data pass the input gate which takes the relevant features from the squashed
input data by multiplying them with a sigmoid function. This function maps
the relevant features the range 0 to 1. If the value is 0, the network removes
the feature, otherwise, the feature pass through the network. The next step
is to decide how much memory we need to store in the cell state. Then a
tanh layer creates a new vector of new candidate values. The memory gate
takes the information stored in the previous state and adds it to the input
gate. Since the memory operator is addition instead of multiplication, LSTM
avoids the vanishing problem. Moreover, the forget data decides which state
of information needs to be memorized or forgotten based sigmoid function.
Finally, the output gate decides what the algorithm is going to output based
sigmoid function. Then, we put the cell state through tanh layer to push the
values to between -1 and 1 and multiply it by the output of the sigmoid gate,
so that we only output the parts we are interested in. In the last layer, the
output is MW. These operations are described as Eqs. (4-9).

c̃(t) = tanh(Wc [a(t−1) , x(t) ] + bc ) (4)

(t) (t−1) (t)
i = σ(Wi [a , x ] + bi ) (5)
f (t) = σ(Wf [a(t−1) , x(t) ] + bf ) (6)
o(t) = σ(Wo [a(t−1) , x(t) ] + bo ) (7)
(t) (t) (t) (t) (t−1)
c =i ∗ c̃ +f ∗c (8)
(t) (t) (t)
a =o ∗ tanh(c ) (9)
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 11

4 Experiment
4.1 Data process
The earthquake information of the regions AETA decting is intercepted from
January 1, 2017 to December 31, 2022. AETA is a multi-component seismic
monitoring system and it is co-developed by the IMS Laboratory of Peking
University and our research group School of Computer Science of Wuhan
University for earthquake monitoring and prediction. AETA system mainly
collects two categories of data, including electromagnetic and acoustic. It has
the advantages of strong system stability, high environmental adaptability,
and strong anti-interference ability [32]. To better verify earthquake predic-
tion, the dataset consists four regions: DS1((98◦ E, 34◦ N), (101◦ E, 30◦ N)),
DS2((101◦ E, 34◦ N), (104◦ E, 30◦ N)), DS3((98◦ E, 30◦ N), (101◦ E, 26◦ N)),
DS4((101◦ E, 30◦ N),(104◦ E, 26◦ N)). The training data starts on January 1,
2017 and ends on February 1, 2022, and the testing data starts on February 1,
2022 and ends on December 30, 2022. In AETA, electromagnetic and acoustic
signals are detected every 10 minutes, so, we selected more than 20000 samples
in the four years only at one station. Each catalog contains a list of records:
time of earthquake, 51 electromagnetic features, and 44 acoustic features. MW
data is maintained on the China Earthquake Network Center (CENC). Their
catalog is available over the internet at http://www.ceic.ac.cn/. Then, we
merge the two tables over the time of earthquake. Table 1 shows the partial
earthquake data of DS2 after merging.

Table 1 Partial earthquake data features of DS2 after merging

Day MW magn@var magn@power magn@skew sound@var sound@power sound@skew

1 0.9 107.685 107.745 -0.003 0.00000248 0.00000248 0.284

2 2.1 22.804 22.809 0.010 0.0000019 0.0000019 0.291
3 0.7 63.482 65.360 -0.047 0.0000019 0.0000019 0.281
... ... ... ... ... ... ... ...
1078 0.9 1.930 1.930 -0.003 0.000320286 0.000320286 0.305
1079 0.2 1.615 1.615 -0.005 0.000319263 0.000319263 0.303
1080 1 1.608 1.608 -0.004 0.00032357 0.00032357 0.300

After merging the dataset, we fuzze multiple pieces of data from one day
into one piece of data. The specific operation consists of two steps: selecting the
different period of catalogs as the representative data and selecting the maxi-
mum of each feature and MW. In this experiment, we choose the timeP eriods
including: 0:00-8:00, 0:00-12:00, and 0:00-24:00. Then, we transform time-series
data into supervised learning data based on MW and the time step is one day.
Fig. 4 shows the kurt of sound which is one of the 95 features.
Fig. 5 shows the relationship between partial different features. From Fig.
5(b), we found the electromagnetic absolute mean value and electromagnetic
absolute maximum 5% position have a strong correlation. Not only these two
features, there are also many characteristics that correlate with each other.
 Springer Nature 2021 LATEX template

12 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

Fig. 4 The wave of the kurt of sound of DS2

Hence, these strong correlation features need to be rejected to decrease time

complexity.

4.2 Prediction indicators

To evaluate the performance of our proposed earthquake prediction model, we
choose various regression evaluation indicators: mean absolute error (MAE),
mean squared error (MSE), RMSE, and R square (R2 ). These indicators are
calculated as Eqs. (10-12).
n
1X
M AE(y, ŷ) = kyi − ŷi k (10)
n i=1
n
1X
M SE(y, ŷ) = k yi − ŷ k22 (11)
n i=1
Xn
k yi − yˆi k22
i=0
R2 (y, ŷ) = 1 − n (12)
X
k yi − y k22
i=0

4.3 Baseline methods

The performance of our proposed EGA-LSTM is compared with other four
ML approaches: RF [17], Adaboost [18], LR [33], and SVR [24]. Moreover,
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 13

(a) (b)

(c) (d)

Fig. 5 Analysis of partial features of DS2

we choose two types of GA: SGA, SteadyGA and three different mutation
strategies of DEs: operation vector is the best vector and the crossover opera-
tion is a random crossover (DE best 1 b), operation vector is the best vector
and the crossover operation is linear order crossover (DE best 1 L), operation
vector is a random vector and the crossover operation is random operation
(DE rand 1 b).
Six EAs adopt binary encoding and set the dimension of individuals to
95 corresponding features. 1 means the corresponding dimension’s feature is
selected, otherwise without being selected. The number of individuals is set
to 10 and the maximum of iteration is set to 20. The threshold of stagnation
is set to 0.000001 and the maximum evolutionary stagnation counter is set to
10. In GAs, the crossover rate and mutation rate are set to 0.7 and 0.01. In
DEs, the scaling factor and crossbreeding rate are both set to 0.5.
The RF is combined with 100 decision trees. Other parameters are set
as follows: max depth = 10, min samples split=2, min samples leaf=1. The
Adaboost’s base learner is decision trees, the number of boosting is 100, and the
learning rate is set to 1 and the boosting algorithm is based on the probability
of prediction error.
 Springer Nature 2021 LATEX template

14 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

The LR’s loss function adopts the least square method. In SVR, the ker-
nel function adopts the radial basis function, the kernel factor is set to the
reciprocal of the product of the number of features and the variance of the
eigenvector and penalty parameter is set to 1.
The LSTM in our experiment has 100 hidden layers, one output layer, and
the activation functions. The learning rate is set to 0.00045, the batch-size is
set to 32 and the number of training epochs are set to 120 epochs. The initial
weights are random values and the bias is set to 0.

4.4 Result and analysis

This section will demonstrate the result in two subsections. In section 4.4.1, the
fitness and RMSE in different periods and EAs are demonstrated. In section
4.4.2, EGA-LSTM is compared with other nine methods based on four metrics.
In section 4.4.3, two non-parametric tests were used to prove that EGA-LSTM
is different other nine algorithms and EGA-LSTM is superior to LSTM in
different datasets.

4.4.1 Different periods and EAs

To find the suitable parameters, we conducted a multi-group ablation exper-
iment with DS2. Tables 2, 3, and 4 refer to the result for time period
(timeP eriods) of 0:00-8:00, 0:00-12:00, and 0:00-24:00, respectively. Fitness is
calculated by Eq. (1). Fig. 6 shows the process of decreasing in EGA searching.
Table 2 demonstrates the result for the timeP eriod of 0:00-8:00, it is worth
noting that EGA performs well than other EAs in most groups. Especially,
in the group of ωa = 1, ωF = 0.8, RMSE is the lowest of all EAs which is a
desirable property and in predicting MW. The difference in indicators of F ,
RM SE, and F itness is markable, producing a better difference greater than
0.03 with the second Fitness (DE rand 1 b) of RMSE and greater than 0.08 of
Fitness. Meanwhile, the number of selected features is only 38 which is lower
than 95 (without selecting). Table 3-4 describe timeP eriods of 0:00-8:00 and
0:00-24:00, respectively, we found RMSE stays between 0.103 and 0.110 and
F decreases in different degrees.
By combining the three tables, we take a comprehensive view of F and
RMSE. Along with lower the number of selected features, RMSE still performs
well. Thus, we conclude that during the period of 0:00-8:00, EGA performs
most stable and suitable. In Fig. 5(b), we found that there is a strong correla-
tion between the electromagnetic absolute mean value and the electromagnetic
absolute maximum 5% position. To increase predicting accuracy and decrease
time complexity, one of these two features needs to be removed. And the best
group’s result shows the target reached.
Fig. 7 shows the trend of RMSE and Fitness scores as ωF increased. With
ωF increasing, the penalties of F . Therefore, the number of selected features
decreased and the trend of Fitness increased. However, the change of RMSE
is not obvious. It always fluctuates in a low range.
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 15

Table 2 Fitness and RMSE in different EAs, parameters, and timeP eriod (0:00-8:00) in
DS2

type of EA ωa = 1, ωF = 0 ωa = 1, ωF = 0.2 ωa = 1, ωF = 0.5 ωa = 1, ωF = 0.8

F RMSE Fitness F RMSE Fitness F RMSE Fitness F RMSE Fitness
EGA 55 0.103 0.103 39 0.104 0.184 39 0.104 0.299 38 0.101 0.417
SGA 49 0.108 0.108 38 0.108 0.190 37 0.105 0.310 37 0.104 0.424
steadyGA 56 0.104 0.104 39 0.103 0.185 38 0.107 0.307 38 0.106 0.426
DE best 1 b 72 0.106 0.106 40 0.105 0.189 33 0.109 0.282 44 0.106 0.476
DE best 1 L 67 0.110 0.110 49 0.104 0.207 38 0.105 0.305 39 0.103 0.431794
DE rand 1 b 63 0.110 0.110 45 0.103 0.198 38 0.104 0.304 38 0.105 0.425

Table 3 Fitness and RMSE in different EAs, parameters, and timeP eriod (0:00-12:00) in
DS2

type of EA ωa = 1, ωF = 0 ωa = 1, ωF = 0.2 ωa = 1, ωF = 0.5 ωa = 1, ωF = 0.8

F RMSE Fitness F RMSE Fitness F RMSE Fitness F RMSE Fitness
EGA 49 0.106 0.106 59 0.107 0.252 39 0.106 0.311 38 0.108 0.722
SGA 43 0.108 0.108 71 0.107 0.256 39 0.107 0.312 39 0.107 0.697
steadyGA 49 0.107 0.107 49 0.111 0.215 43 0.105 0.331 43 0.105 0.636
DE best 1 b 56 0.107 0.107 76 0.104 0.264 43 0.106 0.333 43 0.107 0.722
DE best 1 L 48 0.109 0.109 54 0.106 0.220 43 0.106 0.332 42 0.105 0.619
DE rand 1 b 57 0.110 0.110 55 0.108 0.2244 39 0.105 0.310 39 0.106 0.737

Table 4 Fitness and RMSE in different EAs, parameters, and timeP eriod (0:00-24:00) in
DS2

type of EA ωa = 1, ωF = 0 ωa = 1, ωF = 0.2 ωa = 1, ωF = 0.5 ωa = 1, ωF = 0.8

F RMSE Fitness F RMSE Fitness F RMSE Fitness F RMSE Fitness
EGA 50 0.105 0.105 36 0.104 0.184 37 0.107 0.318 38 0.106 0.409
SGA 63 0.105 0.105 38 0.106 0.181 40 0.104 0.299 36 0.108 0.428
steadyGA 58 0.108 0.108 44 0.112 0.204 42 0.106 0.327 43 0.107 0.469
DE best 1 b 55 0.103 0.103 39 0.105 0.187 38 0.108 0.308 35 0.103 0.397
DE best 1 L 56 0.106 0.106 43 0.104 0.194 31 0.107 0.270 39 0.104 0.432
DE rand 1 b 57 0.105 0.105 45 0.109 0.203 43 0.103 0.329 40 0.106 0.443

4.4.2 Comparisons among ML Methods

To verify the performance of EGA-LSTM, we compare it with some state-of-
the-art ML methods: RF, Adaboost, LR, SVR, and LSTM.
Table 5 demonstrates the effectiveness between EGA-LSTM and other nine
approaches at timeperiod of 0:00-8:00 at the region of DS1. It can be found
that EGA-LSTM obtains better performance on all indicators. Meanwhile,
all the approaches perform badly with the indicator R2 . Totally speaking,
EGA-LSTM outperforms other methods.
Referring to Table 6, we can find that for the MAE, MSE, RMSE and R2 ,
EGA-LSTM performs best among the four predictors. For RMSE, it generates
a difference greater than 0.011 units with the second best predictor (EGA-
SVR). Therefore, EGA-LSTM is an appropriate method to predict MW. The
predictor with the worst result is EGA-LR, which obtains much worse per-
formance on the four indicators than other four approaches. That means the
time sequence between seismic and precursory features cannot be ignored. It
is worth noting that EGA-LSTM’s performance for DS2 is the best among the
four regions.
 Springer Nature 2021 LATEX template

16 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

Fig. 6 The process of decreasing in EGA searching

With the reference to DS3(see Table 7), it is worth noting that LSTM
without EGA perform best than EGA-LSTM. However EGA-LSTM’s results
still perform superior than others for the rest indicators. Intuitively, the dif-
ference between EGA-LSTM and other algorithms is not obvious, especially
EGA-SVR. More statistical comparison details can be seen in section 4.5.
Particular results for DS4 are shown in Table 8. EGA-LSTM perform better
than other nine methods for four metrics.
From a joint analysis of the four tables, a conclusion can be easily con-
cluded. For the four indicators, EGA-LSTM is the best one of them. Therefore,
it can be concluded that EGA-LSTM is the most precise and stable algorithms
for all datasets.
Fig. 8 shows the fitting curves of EGA-LSTM. Since we choose the suitable
parameter and loss function, this model is not overfitting and underfitting.
The real MW and predicted MW on different methods are shown on Fig. 9,
and the result demonstrates that all methods perform well in the range of a
low MW. However, the high MW can rarely be predicted accurately.

Table 5 Comparison of effectiveness between EGA-LSTM and other approaches in DS1

MAE MSE RMSE R2

RF 0.146422 0.033577 0.183239 -0.057718
Adaboost 0.149681 0.033896 0.184108 -0.067774
LR 0.148834 0.035783 0.189163 -0.127214
SVR 0.144821 0.034977 0.187022 -0.101843
LSTM 0.1469 0.03385 0.183984 -0.06633
EGA-RF 0.147592 0.034421 0.18553 -0.084325
EGA-Adaboost 0.146576 0.033112 0.181968 -0.043088
EGA-LR 0.206224 0.06769 0.260173 -1.132344
EGA-SVR 0.14293 0.032739 0.180938 -0.031317
EGA-LSTM 0.140621 0.031833 0.178418 −0.002788
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 17

(a) EGA (b) SGA

(c) steadyGA (d) DE best 1 b

(e) DE best 1 L (f) DE rand 1 b

Fig. 7 The trend of RMSE and Fitness score as ωF

4.5 Non-parametric tests

Since some of the algorithms used were deterministic, we adopt hypothesis
tests. To prove the difference of multiple algorithms, we test the performance
of multiple algorithms on multiple datasets. Earthquake MW datasets can be
used Kruskal-Wallis H test which is a kind of non-parametric test because
they did not assume a normal distribution. Therefore, Kruskal-Wallis H test
was used to determine whether EGA-LSTM’s prediction accuracy significantly
different from other algorithms on datasets DS1, DS2, DS3, and DS4. We set
the confidence level α 0.05. Finally, we designed that the distribution function
of RMSE for group i follows the form Fi (RM SE) = F (RM SE − µi ), where
 Springer Nature 2021 LATEX template

18 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

Table 6 Comparison of effectiveness between EGA-LSTM and other approaches in DS2

MAE MSE RMSE R2

RF 0.090146 0.012795 0.113116 -0.241414
Adaboost 0.085846 0.011548 0.107461 -0.120391
LR 0.085367 0.012644 0.112446 -0.226753
SVR 0.079376 0.010625 0.103077 -0.030835
LSTM 0.080103 0.010495 0.102447 -0.018286
EGA-RF 0.080753 0.010902 0.104411 -0.0577
EGA-Adaboost 0.089556 0.012042 0.109736 -0.168331
EGA-LR 0.145741 0.029862 0.172806 -1.897247
EGA-SVR 0.078192 0.010457 0.102259 -0.014552
EGA-LSTM 0.077376 0.010226 0.101125 0.007829

Table 7 Comparison of effectiveness between EGA-LSTM and other approaches in DS3

MAE MSE RMSE R2

RF 0.13752 0.026332 0.162271 -0.17191
Adaboost 0.138929 0.02688 0.163952 -0.196316
LR 0.152271 0.037694 0.194151 -0.677603
SVR 0.128642 0.023442 0.153108 -0.0433
LSTM 0.121392 0.023618 0.153683 -0.051143
EGA-RF 0.139164 0.027206 0.164944 -0.210831
EGA-Adaboost 0.142054 0.028007 0.167354 -0.246474
EGA-LR 0.160076 0.038701 0.196726 -0.722408
EGA-SVR 0.121748 0.022916 0.15138 -0.019885
EGA-LSTM 0.122361 0.022642 0.150472 −0.007679

µi represents the RMSE of a specific algorithm in our experiments. We set the

null hypothesis(H0 ) and alternative hypothesis(H1 ) as Eqs. (13-14):

H0 : µ1 = µ2 = ... = µ10 (13)

H1 : not all µi are equal (14)

The p-value and statistical quantities of Kruskal-Wallis H test from the
experimental results was 0.0005 and 29.30. The p-value is less than 0.05 which
indicting that we can reject the H0 and receive H1 (that indicts EGA-LSTM
is different from other algorithms).
In order to verify the effectiveness of EGA-LSTM versus LSTM, another
non-parametric statistical test, the Wilcoxon signed-rank test, was conducted
for four different regions between the proposed EGA-LSTM and LSTM. The
hypotheses are list as Eqs. (15-16):

H0 : µLST M <µEGA−LST M (15)

H1 : µLST M ≥ µEGA−LST M (16)

 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 19

Table 8 Comparison of effectiveness between EGA-LSTM and other approaches in DS4

MAE MSE RMSE R2

RF 0.109322 0.020374 0.142739 -0.184501
Adaboost 0.107933 0.01903 0.137949 -0.106342
LR 0.107009 0.020991 0.144883 -0.220349
SVR 0.103363 0.018632 0.136498 -0.083187
LSTM 0.100836 0.017347 0.131709 -0.00851
EGA-RF 0.106098 0.019892 0.141041 -0.156479
EGA-Adaboost 0.105577 0.01833 0.13539 -0.065664
EGA-LR 0.11569 0.023812 0.154312 -0.384367
EGA-SVR 0.100364 0.017603 0.132678 -0.023402
EGA-LSTM 0.099404 0.017231 0.131267 −0.001748

Fig. 8 Plot of epoch and loss of the LSTM model

The µLST M and µEGA−LST M represent the RMSE of the two models. H0
means LSTM perform better than EGA-LSTM and H1 the opposite. The
description and quartile statistics are presented in Table 9 for the two methods.
Sub-Table 9(a) shows the basic test statistics, such as maximum, minimum,
standard deviation, and other values for the pairwise comparison of EGA-
LSTM and LSTM based on each evaluation indicators separately. It can be
observed from Table 11 for 20 independent and replicate experiments, that
RMSE of EGA-LSTM is better than the opposed to its counterpart LSTM.
In sub-Table 9(b) of Wilcoxon signed-rank, positive ranks, negative ranks
are presented for each region of measures corresponding to EGA-LSTM and
LSTM. The Wilcoxon signed-rank shows that the results from different regions
are different. Checking the Wilcoxon signed-rank’s boundary table, the critical
value in case of one-tailed test with confidence level of 0.01 is 43, and in case of
the confidence level of 0.05 is 60. In DS1, the test statistic is 28, therefore we
reject the null hypothesis with 99% certainty that EGA-LSTM is superior to
 Springer Nature 2021 LATEX template

20 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

Fig. 9 Comparison between real and predicted MW

LSTM. DS4 is the same as DS1. The test statistic of DS2 is 58 with is greater
than 43, we can not reject the null hypothesis with the confidence level of 0.01.
However, 58 is less than 60 which is the critical value with confidence level of
0.05. DS3 is the same az DS2. In summary, they all indicate strong evidence
for the null hypothesis, that EGA-LSTM is superior than the standard LSTM.
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 21

Table 9 Results of Wilcoxon signed-rank test in terms of statistical measures

N Mean Std. Min Max Percentiles
25th 50th 75th
a) Descriptive Statistics
1) DS1
LSTM-RMSE 20 0.181 4.81E-6 0.178 0.186 0.179 0.181 0.182
EGA-LSTM-RMSE 20 0.179 1.40E-6 0.177 0.182 0.178 0.179 0.180
2) DS2
LSTM-RMSE 20 0.105 1.50E-5 0.100 0.114 0.102 0.103 0.108
EGA-LSTM-RMSE 20 0.104 2.30E-5 0.100 0.11 0.101 0.102 0.107
3) DS3
LSTM-RMSE 20 0.157 6.17E-6 0.153 0.162 0.155 0.156 0.158
EGA-LSTM-RMSE 20 0.155 2.04E-6 0.152 0.157 0.154 0.155 0.156
4) DS4
LSTM-RMSE 20 0.132 0.10E-6 0.130 0.134 0.131 0.132 0.132
EGA-LSTM-RMSE 20 0.131 0 0.130 0.133 0.131 0.131 0.131
N Mean Rank Sum of Ranks
b) Wilcoxon Signed-Ranks Test
1) DS1
RMSE: LSTM - EGA-LSTM Positive Ranks 15a 12.13 182 a. RMSE: LSTM > EGA-LSTM
Negative Ranks 5b 5.60 28 b. RMSE: LSTM < EGA-LSTM
Ties 0c c. RMSE: LSTM = EGA-LSTM
Total 20
2) DS2
RMSE: LSTM - EGA-LSTM Positive Ranks 13a 10.62 138
Negative Ranks 7b 10.29 72
Ties 0c
Total 20
3) DS3
RMSE: LSTM - EGA-LSTM Positive Ranks 13a 12.85 167
Negative Ranks 7b 6.14 43
Ties 0c
Total 20
4) DS4
RMSE: LSTM - EGA-LSTM Positive Ranks 16a 11.50 184
Negative Ranks 4b 6.50 26
Ties 0c
Total 20
RMSE: LSTM - EGA-LSTM DS1 DS2 DS3 DS4
c) Testing
Test statistic 28 58 47 26

5 Conclusion and Future Works

Aiming at solving the problem that the redundant features and time series
properties hinder the development of earthquake magnitude prediction mod-
els, we propose an EGA-LSTM for time series earthquake prediction. First,
the acoustic and electromagnetics data of AETA system we developed are
fused and preprocessed by EGA to find the strong correlation indicators.
Second, since the EGA has the advantages in searching optimal feature sub-
set, we adopt it to selected features. Then, LSTM is implemented to execute
magnitude prediction with the selected features, to process time series and
complex data. Specifically, we chose RMSE of LSTM and the ratio of selected
features as the fitness components of EGA. Finally, we test the proposed EGA-
LSTM on the AETA data of Sichuan and Yunnan province. Experimental
results demonstrate that all the methods can get the best performance when
timeP eriod = 0 : 00 − 8 : 00, ωa = 1, and ωF = 0.8. Moreover, our proposed
EGA-LSTM obtains satisfying performance than state-of-the-art approaches
on the evaluation indicators MAE, MSE, RMSE, and R2 .
However, due to the data of medium and large earthquakes belonging to
small samples, in order to be able to predict medium and large earthquakes,
the model proposed in this study needs to be improved for the data in AETA.
Our future work will focus on how to process the time series of small samples
to obtain an effective and usable magnitude prediction model, which is suitable
for medium and large earthquakes. Apart from that, the theoretical analysis
 Springer Nature 2021 LATEX template

22 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

and more complicated earthquake prediction scenarios of our method can also
be part of future work.

6 Ethical Approval
Not applicable.

7 Competing interests
Not applicable.

8 Authors’ contributions
Zhiwei Ye and Wuyang Lan and Wen Zhou: Wrote the main manuscript
text. Qiyi He: Supervision, Writing- Reviewing and Editing, Funding acquisi-
tion. Liang Hong and Xinguo Yu and Yunxuan Gao: Reviewing. All authors
reviewed the manuscript.

9 Availability of data and materials

The datasets generated during and/or analysed during the current study are
available from the corresponding author on reasonable request.

10 Funding
The authors want to thank NSFC- http://www.nsfc.gov.cn/ for the support
through Grants Number 61877045 and 62202147, and Fundamental Research
Project of Shenzhen Science and Technology Program for the support through
Grants Number JCYJ20160428153956266, and Research project of the Natu-
ral Resources Department of Hubei Province for the support through Grants
Number ZRZY2023KJ13.

References
[1] Bank, W., Nations, U.: Natural Hazards, Unnatural Disasters: the
Economics of Effective Prevention. The World Bank, ??? (2010)

[2] Adeli, H., Panakkat, A.: A probabilistic neural network for earthquake
magnitude prediction. Neural networks 22(7), 1018–1024 (2009)

[3] Asim, K.M., Moustafa, S.S., Niaz, I.A., Elawadi, E.A., Iqbal, T., Martı́nez-
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)
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 23

[4] Berhich, A., Belouadha, F.-Z., Kabbaj, M.I.: Lstm-based earthquake pre-
diction: enhanced time feature and data representation. International
Journal of High Performance Systems Architecture 10(1), 1–11 (2021)

[5] Cai, Y., Shyu, M.-L., Tu, Y.-X., Teng, Y.-T., Hu, X.-X.: Anomaly detec-
tion of earthquake precursor data using long short-term memory networks.
Applied Geophysics 16, 257–266 (2019)

[6] Kadam, V.J., Yadav, S.S., Jadhav, S.M.: Soft-margin svm incorporating
feature selection using improved elitist ga for arrhythmia classification.
In: Intelligent Systems Design and Applications: 18th International Con-
ference on Intelligent Systems Design and Applications (ISDA 2018) Held
in Vellore, India, December 6-8, 2018, Volume 2, pp. 965–976 (2020).
Springer

[7] John, G.H., Kohavi, R., Pfleger, K.: Irrelevant features and the subset
selection problem. In: Machine Learning Proceedings 1994, pp. 121–129.
Elsevier, ??? (1994)

[8] Martı́nez-Álvarez, F., Reyes, J., Morales-Esteban, A., Rubio-Escudero, C.:

Determining the best set of seismicity indicators to predict earthquakes.
two case studies: Chile and the iberian peninsula. Knowledge-Based
Systems 50, 198–210 (2013)

[9] Roiz-Pagador, J., Chacon-Maldonado, A., Ruiz, R., Asencio-Cortes, G.:

Earthquake prediction in california using feature selection techniques.
In: International Workshop on Soft Computing Models in Industrial and
Environmental Applications, pp. 728–738 (2021). Springer

[10] Chen, Y., Zhang, J., He, J.: Research on application of earthquake pre-
diction based on chaos theory. In: 2010 International Conference on
Intelligent Computing and Integrated Systems, pp. 753–756 (2010). IEEE

[11] Cekim, H.O., Tekin, S., Özel, G.: Prediction of the earthquake magni-
tude by time series methods along the east anatolian fault, turkey. Earth
Science Informatics 14(3), 1339–1348 (2021)

[12] Panakkat, A., Adeli, H.: Neural network models for earthquake magnitude
prediction using multiple seismicity indicators. International journal of
neural systems 17(01), 13–33 (2007)

[13] Asim, K., Martı́nez-Álvarez, F., Basit, A., Iqbal, T.: Earthquake magni-
tude prediction in hindukush region using machine learning techniques.
Natural Hazards 85(1), 471–486 (2017)

[14] Chanda, S., Raghucharan, M., Reddy, K.K., Chaudhari, V., Somala, S.N.:
Duration prediction of chilean strong motion data using machine learning.
 Springer Nature 2021 LATEX template

24 Elite GA-based Feature Selection of LSTM for Earthquake Prediction

Journal of South American Earth Sciences 109, 103253 (2021)

[15] Shah, H., Ghazali, R.: Prediction of earthquake magnitude by an improved

abc-mlp. In: 2011 Developments in E-systems Engineering, pp. 312–317
(2011). IEEE

[16] Muhammad, A., Külahcı, F., Birel, S.: Investigating radon and TEC
anomalies relative to earthquakes via AI models. J. Atmos. Sol. Terr.
Phys. 245(106037), 106037 (2023)

[17] Yang, F., Kefalas, M., Koch, M., Kononova, A.V., Qiao, Y., Bäck, T.:
Auto-rep: An automated regression pipeline approach for high-efficiency
earthquake prediction using lanl data. In: 2022 14th International Confer-
ence on Computer and Automation Engineering (ICCAE), pp. 127–134
(2022). IEEE

[18] Asim, K.M., Idris, A., Iqbal, T., Martı́nez-Álvarez, F.: Seismic indica-
tors based earthquake predictor system using genetic programming and
adaboost classification. Soil Dynamics and Earthquake Engineering 111,
1–7 (2018)

[19] Zhou, W., Liang, Y., Dong, H., Tan, C., Xiao, Z., Liu, W.: A numerical dif-
ferentiation based dendritic cell model. In: 2017 IEEE 29th International
Conference on Tools with Artificial Intelligence (ICTAI), pp. 1092–1098
(2017). IEEE

[20] Zhou, W., Dong, H., Liang, Y.: The deterministic dendritic cell algo-
rithm with haskell in earthquake magnitude prediction. Earth Science
Informatics 13(2), 447–457 (2020)

[21] Zhou, W., Zhang, K., Ming, Z., Chen, J., Liang, Y.: Immune optimization
inspired artificial natural killer cell earthquake prediction method. The
Journal of Supercomputing 2022, 1–23 (2022)

[22] Zhou, W., Liang, Y., Wang, X., Ming, Z., Xiao, Z., Fan, X.: Introduc-
ing macrophages to artificial immune systems for earthquake prediction.
Applied Soft Computing 122, 108822 (2022)

[23] Moustra, M., Avraamides, M., Christodoulou, C.: Artificial neural net-
works for earthquake prediction using time series magnitude data or
seismic electric signals. Expert systems with applications 38(12), 15032–
15039 (2011)

[24] Asim, K.M., Idris, A., Iqbal, T., Martı́nez-Álvarez, F.: Earthquake pre-
diction model using support vector regressor and hybrid neural networks.
PloS one 13(7), 0199004 (2018)
 Springer Nature 2021 LATEX template

Elite GA-based Feature Selection of LSTM for Earthquake Prediction 25

[25] 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. Multimedia Tools
and Applications 80(18), 28419–28438 (2021)

[26] Draz, M.U., Shah, M., Jamjareegulgarn, P., Shahzad, R., Hasan, A.M.,
Ghamry, N.A.: Deep machine learning based possible atmospheric and
ionospheric precursors of the 2021 mw 7.1 japan earthquake. Remote
Sensing 15(7) (2023). https://doi.org/10.3390/rs15071904

[27] Berhich, A., Belouadha, F.-Z., Kabbaj, M.I.: Lstm-based models for earth-
quake prediction. In: Proceedings of the 3rd International Conference on
Networking, Information Systems & Security, pp. 1–7 (2020)

[28] Berhich, A., Belouadha, F.-Z., Kabbaj, M.I.: An attention-based LSTM

network for large earthquake prediction. Soil Dyn. Earthq. Eng.
165(107663), 107663 (2023)

[29] Kavianpour, P., Kavianpour, M., Jahani, E., Ramezani, A.: A cnn-bilstm
model with attention mechanism for earthquake prediction. arXiv preprint
arXiv:2112.13444 (2021)

[30] Jh, H.: Adaptation in natural and artificial systems. Ann Arbor (1975)

[31] Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural com-
putation 9(8), 1735–1780 (1997)

[32] Wanga, J., Yong, S., et al.: An aeta electromagnetic disturbance anomaly
extraction method based on sample entropy. In: 2021 IEEE 5th Advanced
Information Technology, Electronic and Automation Control Conference
(IAEAC), vol. 5, pp. 2265–2269 (2021). IEEE

[33] Adeli, H., Panakkat, A.: A probabilistic neural network for earthquake
magnitude prediction. Neural Networks 22(7), 1018–1024 (2009)
Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.

EQTransformermaster.zip
areafeature.rar