Eurasian Economic Review (2023) 13:803–830

https://doi.org/10.1007/s40822-023-00243-x

ORIGINAL PAPER

Forecasting volatility by using wavelet transform, ARIMA

and GARCH models

Lihki Rubio1 · Adriana Palacio Pinedo1 · Adriana Mejía Castaño1 ·

Filipe Ramos2

Received: 29 March 2023 / Revised: 19 September 2023 / Accepted: 14 October 2023 /

Published online: 10 December 2023
© The Author(s) 2023

Abstract
Forecasting volatility of certain stocks plays an important role for investors as it
allows to quantify associated trading risk and thus make right decisions. This work
explores econometric alternatives for time series forecasting, such as the ARIMA
and GARCH models, which have been widely used in the financial industry. These
techniques have the advantage that training the models does not require high com-
putational cost. To improve predictions obtained from ARIMA, the discrete Fourier
transform is used as ARIMA pre-processing, resulting in the wavelet ARIMA strat-
egy. Due to the linear nature of ARIMA, non-linear patterns in the volatility time
series cannot be captured. To solve this problem, two hybridisation techniques are
proposed, combining wavelet ARIMA and GARCH. The advantage of applying this
methodology is associated with the ability of each to capture linear and non-linear
patterns present in a time series. These two hybridisation techniques are evaluated to
verify which provides better prediction. The volatility time series is associated with
Tesla stock, which has a highly volatile nature and it is of major interest to many
investors today.

Keywords Volatility · ARIMA · GARCH · Wavelet transform · Hybrid model

* Filipe Ramos
frramos@fc.ul.pt
Lihki Rubio
lihkir@uninorte.edu.co
Adriana Palacio Pinedo
mapalacio@uninorte.edu.co
Adriana Mejía Castaño
mejiala@uninorte.edu.co
1
Department of Mathematics and Statistics, Universidad del Norte, Km. 5 vía Puerto Colombia,
081007 Barranquilla, Atlántico, Colombia
2
CEAUL‑Centro de Estatística e Aplicações, Faculdade de Ciências, Universidade de Lisboa,
1749‑016 Lisbon, Portugal

13
Vol.:(0123456789)
804 Eurasian Economic Review (2023) 13:803–830

Abbreviations
ARIMA Autoregressive integrated moving average
GARCH Generalized autoregressive conditional heteroskedasticity
DWT Discrete wavelet transform
IDWT Inverse discrete wavelet transform
Ait  ith approximation component from DWT at the time t
Dit  ith detailed component from DWT at the time t
yt Time series observation and prediction at time t
yt , ̃
̃
hw-ARIMA
t
W-ARIMA prediction obtained by applying ARIMA model to
approximate (Ait ) and detailed (Dit ) components from (DWT)
decomposition
̃
hw-GARCH
t
W-GARCH prediction obtained by applying GARCH model to
approximate (Ait ) and detailed (Dit ) components from (DWT)
decomposition
̃
hw-ARIMA-GARCH
t
W-ARIMA-GARCH prediction obtained by applying ARIMA to
approximate components (Ait ) and GARCH to detailed compo-
nents (Dit ) from (DWT) decomposition
̃
hw-GARCH-ARIMA
t
W-GARCH-ARIMA prediction obtained by applying GARCH to
approximate components (Ait ) and ARIMA to detailed compo-
nents (Dit ) from (DWT) decomposition
̃
hARIMA-GARCH
t
 ARIMA-GARCH prediction obtained by adding ARIMA forecast
from the original time series yt and GARCH prediction for its
residuals
̃
hGARCH-ARIMA
t
 GARCH-ARIMA prediction obtained by adding GARCH forecast
from the original time series yt and ARIMA prediction for its
residuals
̃
hw-ARIMA-SV
t
Prediction obtained by applying ARIMA to approximate com-
ponents (Ait ) and SV to detailed components (Dit ) from (DWT)
decomposition

1 Introduction

With the creation of multiple financial indices and the increasing number of com-
panies listing their shares on stock exchanges, the demand for forecasting models to
quantify inherent risk in each of such instruments has grown considerably. Financial
firms and traders are frequently concerned about the risks linked to the increasing
volatility of stocks, such as Tesla, Inc. (TSLA) listed in Nasdaq, whose volatility
will be predicted in this paper for illustrative purposes due to high volatility pre-
sent in this stock. Tesla shares rose strongly at the time of the COVID-19 pandemic.
However, the month-on-month decrease in the stock is deeper than the Represent the
Bitcoin symbol (BTC) price drop. The problem is speculated to be associated with
the CEO’s sale of a portion of his stake to help fund new acquisitions. Investors are
now questioning whether the sale has been finalized, which is causing high volatility
in the stock.

13
Eurasian Economic Review (2023) 13:803–830 805

On the other hand, accurate volatility prediction is a crucial part of risk manage-
ment, most notably in asset allocation to diverse investment portfolios to adequately
hedge the underlying risk. High volatility indicates that either the market is at risk
or that securities values are unreliable and that capital markets are not performing
well enough. Gaining a more accurate understanding of volatility, predicting it with
accuracy and controlling portfolio exposure and its impact are essential to effective
trading.
Different authors have studied volatility forecasting using statistical and machine
learning models. For example, some of them have opted to use wavelet transform
support vector machine (WSVM) to outperformed solutions obtained from (SVM)
(Tang et al., 2009). Comparative studies on different versions of the Generalized
Autoregressive Conditional Heteroskedasticity (GARCH) model, such as Cholesky
Generalized Autoregressive Conditional Heteroskedasticity (CGARCH), Inte-
grated Generalized Autoregressive Conditional Heteroskedasticity (IGARCH) and
Fractionally Integrated Generalized Autoregressive Conditional Heteroskedasticity
(FIGARCH), to detect which one delivers the best performance have been devel-
oped (Kang et al., 2009). Vector autoregressive models such as the VARIMA and
VGARCH model along with aggregated causality factors have also been used in
volatility forecasting (Hafner, 2009). The use of particle filtering (PF) techniques
has also been applied in volatility forecasting for high frequency data, which con-
tains patterns difficult to capture by classical statistical models (Gaoyu et al., 2009).
In addition, techniques based on artificial neural networks together with classical
time series models have been used in volatility forecasting and electricity demand to
improve prediction accuracy (Hyup Roh, 2007; Luzia et al., 2023).
The main problem with volatility prediction lies in the fact that strong nonlin-
ear patterns may be present in the associated time series. Therefore, linear models
such as ARIMA or Exponential Smoothing (ES), which are widely used in statistics,
for example to predict earnings before interest, taxes, depreciation and amortization
(EBITA) index, mid-long-term electric energy consumption and exchange-rates (de
Oliveira & Cyrino Oliveira, 2018; Maria & Eva, n.d.; Rubio et al., 2021) others are
unable to predict volatility with high efficiency. These types of models are defined
in terms of linear filters considering a finite number of historical data and residuals,
to fit linear patterns in a time series. An alternative way to improve goodness-of-fit
is to use moving average or discrete Fourier transform as pre-processing (Al Wadi
et al., 2010; Alshammari et al., 2020).
Another option is to use machine learning (ML) or deep learning (DL) techniques
which are able to capture strong fluctuations in a time series or have important prop-
erties associated with long term time dependencies, such as support vector regres-
sion (SVR) and long short-term memory (LSTM) models, respectively (Bathla,
2020; Chniti et al., 2017; Guo et al., 2019). A principal concern associated with the
use of these approaches is the high computational cost required to train these mod-
els. GARCH stochastic processes have shown great effectiveness in predicting time
series with implicit heteroscedasticity, such as volatility, and their evaluation does
not require large computational resources. Therefore, the use of this technique on its
own or in combination with another method will provide predictions with adequate
scores for volatility forecasting.

13
806 Eurasian Economic Review (2023) 13:803–830

In this paper, a hybridization technique based on the wavelet transform, using

the ARIMA and GARCH models, is proposed. This technique uses selected models
for each of the time series obtained from the decomposition into high and low fre-
quency signals. Time series with high frequency are predicted using GARCH and
those with low frequency are predicted using ARIMA or wavelet ARIMA. In this
way, advantages offered by each technique when predicting time series volatility
are maximized. A comparative study between this type of hybridization and other
related techniques is performed to identify which approach provides the best combi-
nation of these predictive models.
The primary objective of this work is to demonstrate that the model, which incor-
porates wavelet transform, ARIMA, and GARCH, with the application of these
models to both approximate and detailed components, yields superior predictions
compared to other hybrid models utilized for volatility forecasting. To achieve this
goal, various iterations of hybrid models, integrating wavelet transform and the
incorporation of predicted residuals, are assessed. This assessment aims to establish
that the suggested model provides the most favourable outcomes in terms of error
rates and goodness of fit.
Having described the background of this research, the organization of the other
sections shall be as follows: Sect. 2 includes a review of the literature associated
with the problem addressed in this work. Section 3 describes formally predictive
models used for volatility forecasting and associated important definitions. Section 4
introduce data set used for time series volatility estimation, its associated descriptive
analysis. Section 5 and 6 present the methodology proposed for the optimal combi-
nation of ARIMA and GARCH models, experimental results obtained and discus-
sions related. Section 7 present conclusions, future and on-going work.

2 Literature review

As an alternative to ARIMA, wavelet ARIMA model has been used by different

authors for volatility forecasting using diverse approaches. For example, it has been
shown that wavelet ARIMA provides better predictions than ARIMA for forecast-
ing Amman stocks. In this case, ARIMA was applied to each of the Daubechies
wavelets obtained from the decomposition (Al Wadi et al., 2010). Other authors use
for example, maximal overlap discrete wavelet transform (MODWT) and hybrid
ARIMA models with feedforward neural network FNN for volatility forecasting. In
this case, it was found that the hybrid ARIMA-FNN model provides better predic-
tions in terms of goodness-of-fit. ARIMA-FNN uses first ARIMA to fit linear pat-
terns in the time series, then residuals are predicted using the FNN model. In this
fashion, FNN adjusts prediction errors obtained from ARIMA (Xiao et al., 2014).
Although wavelet transform is crucial to improve volatility predictions obtained
from ARIMA, there are some strong fluctuations that cannot be captured due to its
linear structure. Some authors have opted for using machine learning and deep learn-
ing techniques to solve this problem. For example, using ML models such as neural
networks and random forests for volatility prediction of the Dow Jones Industrial
Average index stocks yields considerable predictive gains, since these models can

13
Eurasian Economic Review (2023) 13:803–830 807

effectively capture multiple highly nonlinear patterns in a time series, which cannot
be fitted by models such as heterogeneous autoregressive (HAR) (Christensen et al.,
2021). Support vector machine models dominated the area of machine learning in
the 1990s until 2010, displacing techniques based on artificial neural networks.
After that, the deep learning-based techniques began to dominate, thanks to the cur-
rent availability of computing power, advances in computer architectures, as well as
the accumulation of big data. However, the SVM model has a great advantage over
alternative models, as it is able to capture long-memory nonlinear patterns in time
series such as the volatility of the S&P 500 index, it is additionally effective for high
dimensionality data processing. SVM has been compared with classical models used
in econometrics, for example, multiple versions of the GARCH model, obtaining
comparable or better predictions using SVM than those obtained using the GARCH
models (Gavrishchaka & Banerjee, 2006).
On the other hand, some authors prefer to use deep learning techniques such as
long short-term memory (LSTM) for volatility prediction. It has been found that
these techniques provide similar or better predictions in some cases to those pro-
vided by the support vector regression (SVR) model, which is highly suggested in
the literature for this type of forecasting, as it is effective when it comes to adjusting
marked patterns of heteroskedasticity. The use of LSTM is highlighted by the fact
that it can be trained using graphics cards to reduce computational cost (Liu, 2019).
Recent research has also considered to make use of stacked machine learning mod-
els based on artificial neural networks to forecast volatility of the S&P 500 index. In
this case, predictions are obtained by evaluating under an activation function, fore-
casts obtained through Gradient Descent Boosting (GDC), Random Forest (RF) and
Support Vector Machine (SVM) models, to obtain a single stacked forecast. This
model has proven to be competitive with the classical models used individually
(Ramos-Pérez et al., 2019).
Another alternative used to forecast realized volatility is to apply the hybrid
ARIMA-GARCH model to each signal obtained from MODWT decomposition. The
final forecast is obtained by adding forecasts from each MODWT wave component.
Authors highlight advantages of using MODWT as a pre-processor for hybrid mod-
els such as ARIMA-GARCH to predict volatility of the Saudi Arabia stock market
(Alshammari et al., 2020).
Several successful applications of wavelet transform-based methods has been
used in finance applications. One such example is enhancing stock index prediction
accuracy using a closed recurrent unit (GRU) neural network in conjunction with
adaptive noise decomposition (CEEMDAN-wavelet). The results demonstrate that
the GRU-CEEMDAN-wavelet approach yields lower errors compared to ARIMA
and individual models (Qi et al., 2023). On the other hand, Li and Tang (2020) pro-
posed the WT-FCD-MLGRU model to enhance accuracy of forecasting financial
time series linked to stock indices. This model integrates wavelet transform, filter
cycle decomposition, and multilag neural networks. Empirical analysis reveals that
the WT-FCD-MLGRU model outperform alternative approaches in forecast preci-
sion, showing the lowest error when predicting stock indices, as compared to con-
ventional models as ARIMA and the enhanced SVR machine learning model. Wang
and Guo (2020) introduces the hybrid model DWT-ARIMA-GSXGB. The authors

13
808 Eurasian Economic Review (2023) 13:803–830

use discrete wavelet transform to partition the dataset into approximation and error
components. Subsequently, ARIMA is applied to the approximate partial data, while
the enhanced XGBoost model (GSXGB) is used for the error data. Through experi-
mental comparisons with stock market data, it was found that the DWT-ARIMA-
GSXGB model exhibits lower errors when contrasted with four prediction mod-
els: ARIMA, XGBoost, GSXGB, and DWT-ARIMA-XGBoost. A hybrid model
which integrates the Empirical Wavelet Transform (EWT) with the Improved Bee
Colony Algorithm (ABC), the Extreme Learning Machine (ELM) neural network,
and ARIMA has been proposed (Yu et al., 2020). EWT is employed to decompose,
clean the data and removing noise, making it suitable for forecasting. Subsequently,
the optimized ELM using GPS-EO-ABC and ARIMA are separately applied to gen-
erate diverse prediction outcomes, which are then combined with weighting. The
optimized ELM demonstrated superior accuracy and stability when compared to the
original ELM, ABC-ELM, LSTM, and ANN. This hybrid approach proves to be
effective in not only prediction but also efficient in noise reduction and outlier cor-
rection within the data.
Each of the studies mentioned previously have proposed novel techniques for
volatility forecasting, with a common purpose, to find a technique to get accurate
and trustworthy predictions for this type of data, which has a large amount of noise
and strong behaviour difficult to adjust, due to different external factors and current
market conditions. The literature review did not reveal the existence of a model that
combines wavelet transform, ARIMA, and GARCH for volatility prediction. In this
proposed approach, ARIMA and GARCH are specifically employed for the approxi-
mate and detailed components derived from discrete wavelet transform.

3 Forecasting models

3.1 ARIMA model

ARIMA models, popularized by Box and Jenkins, are a flexible and powerful sta-
tistical tool for predictive modelling with time series data (Asteriou & Hall, 2016).
Mainly, ARIMA models approximate time series future values as a linear function
of past observations and white noise terms. The model consists of three components:
non-stationary differences for stationarity, autoregressive model (AR) and moving
average (MA) model (Montgomery et al., n.d.).
To define non-stationarity, the backshift operator, B is introduced. A time series,
yt , will be called homogeneous non-stationary if it is non-stationary but its first
difference, i.e. wt = yt − yt−1 = (1 − B)yt or d th difference, wt = (1 − B)d yt , yields
a stationary time series. In addition, yt will be called an autoregressive integrated
moving average (ARIMA) process of orders p, d, and q, denoted ARIMA (p, d, q) if
its d th difference yields a stationary process ARMA(p, q). Therefore, an ARIMA (p,
d, q) can be written as:

Φ(B)(1 − B)d yt = 𝛿 + Θ(B)𝜀t , (1)

13
Eurasian Economic Review (2023) 13:803–830 809

where:
p q
∑ ∑
Φ(B) = 1 − 𝜙i Bi , Θ(B) = 1 − 𝜃i Bi (2)
i=1 i=1

are the backshit operator terms in the AR(p) and MA(q) defined as: Φ(B)yt = 𝛿 + 𝜀t
and yt (= 𝜇) + Φ(B)𝜀t , with 𝛿 = 𝜇 − 𝜙𝜇, where μ is the mean and 𝜀t the white noise
with E 𝜀t = 0.
Model orders p, q are determined by the nature of the autocorrelation and partial
autocorrelation functions. The model coefficients are calculated using the maximum
likelihood method (Box et al., 2008). The best model is identified by diagnostic
checks such as the Akaike information criterion (AIC), the Bayes information crite-
rion (BIC) and the Jarque–Bera normality test on the residual error series.

3.2 Wavelet‑based ARIMA

The wavelet-based ARIMA is a composite technique, in which the raw data is

decomposed first using wavelet transformation (WT) and then forecasting models
are applied. The WT decomposes a signal into different time scales. It is defined as a
set of basis functions 𝜓a,b (t) that can be generated by translating and scaling the so-
called mother wavelet (Daubechies, 1992)
� �
1 t−a
𝜓a,b = √ 𝜓 , a > 0, −∞ < b < ∞, (3)
a a

where a is the scale parameter and b the location of the wavelet. There are several
types of WT for which different mothers can be used, in this work we will concen-
trate on the mostly used Discrete Wavelet Transform (DWT).
A DWT is a discrete set of wavelet scales and translations. The DWT is adapted
mainly for samples. Although, this transform decomposes the signal into a multi-
orthogonal set of wavelets. DWT uses a dyadic grid, where the mother wavelet is
scaled by the power two (a = 2j ) and shifted by an integer (b = k2j ), where k is a
location index ranges from 1 to 2−j N ( N number of observations) and j from 0 to J
( J total number of scales). DWT is given by the following equation:
( )
Ψj,k (t) = 2−j∕2 Ψ 2−j t − k , (4)

where the coefficients are calculated according to the following expression:

∞

∫
( ) ( )
Wj,k = W 2j , k2j = 2−j∕2 f (t)Ψ 2−j t − k dt (5)
−∞

The inverse discrete wavelet transform (IDWT) mean is then calculated to re-con-
struct the original signal from the wavelet coefficients Wj,k as follows:

13
810 Eurasian Economic Review (2023) 13:803–830

∞ ∞
∑ ∑
f (t) = Wj,k Ψj,k (t) (6)
j=−∞ k=−∞

There are different mother wavelets in the DWT, for example the Haar wavelet,
the Daubechies wavelet, the orthogonal wavelet, the Symlet wavelet, the Meyer
wavelet and the Coiflets wavelet (Mallat & Peyré, n.d.). The Haar wavelet is consid-
ered the simplest mother wavelet, while the Daubechies wavelet is a set of orthogo-
nal wavelets. In this paper, Daubechies mother wavelets will be used for our first
pre-processing. Symlet and Coiflet are defined as a modified version of Daube-
chies wavelets. On the other hand, Symlet is a Daubechies wavelet with a higher
symmetry.
Multi-Resolution Analysis (MRA) also called pyramid algorithm, is defined by a
hierarchical representation of DWT (Mallat, n.d.). This is based on the decomposi-
tion of the raw data into m levels by translation and convolution of the mother wave-
let using low-pass (LP) and high-pass (HP) filters. Detail (D) and approximation (A)
components are kept using these filters. The signal can be reconstructed through the
sum of the last approximation component and every detail component.
The wavelet transform is employed to eliminate noise from the time series,
thereby enhancing the stability of the data structure. As an example, consider a
time series yt . By performing Daubechies wavelet filtering on yt , two sets of filtered
and decomposed series are generated, the (LP) and (HP) versions, named the i -th
approximate and detailed components at the time t , denoted by Ait and Dit , respec-
tively (see Fig. 1).
Each component can be further filtered to obtain a second level for each one. The
decomposed time series is predicted by applying the according time series model to
each component. For example, the ARIMA model described in Sect. 3.1 can be used
to predict the Dit and Ait components, the final predictive model for the original time
series, denoted by w-ARIMA can be obtained by the following addition.

yw-ARIMA
̃ ̃ ARIMA + D
=A ̃ ARIMA + D
̃ ARIMA (7)
t 2t 2t 1t

where each term of the sum is obtained from the inverse reconstruction of signals
(IDWT) (see Fig. 2). This model uses Daubechies filter as pre-processing to reduce
noise in the time series data and then apply ARIMA to each component, to obtain
higher prediction accuracy than the classical ARIMA.

HP
HP

LP
LP

Fig. 1  Multi-resolution analysis applied to the original time series yt , where Ait represent the approxima-
tion components, and Dit , the detailed. (HP) is the high pass filter and (LP) is the low pass filter

13
Eurasian Economic Review (2023) 13:803–830 811

Fig. 2  Flow chart for Python implementation of partitioned models to forecast volatility using PyWave-
lets, statsmodels.tsa.arima and arch.arch model libraries

3.3 GARCH models

Heteroscedastic conditional autoregressive (ARCH) models first studied by Engle

(1982) have been widely used for high volatility forecasting. These models effec-
tively describe the variance changing over time i.e. heteroscedasticity, by employing
a deterministic mapping based on historical errors. Formally, the ARCH model of
order p ≥ 0 is defined as follows:
p
√ �
yt = ht 𝜀t , ht = 𝛼0 + 𝛼i y2t−i (8)
i=1

The generalized autoregressive conditional heteroskedasticity (GARCH)

model can be considered as a generalization of the ARCH model defined by
Eq. (9). GARCH aggregates linear combination of conditional variance squares
and for orders p ≥ 0 and q ≥ 0 is given by
p q
√ � �
yt = ht 𝜀t , ht = 𝛼0 + 𝛼i y2t−i + 𝛽j h2t−j , (9)
i=1 j=1

where 𝜀t ∼ iid(0, 1), 𝛼0 > 1, 𝛼i ≥ 0, i = 1, 2, … , p, 𝛽j ≥ 0, j = 1, 2, … , q,

13
812 Eurasian Economic Review (2023) 13:803–830

p q
∑ ∑
𝛼i + 𝛽j < 1, (10)
i=1 j=1

and 𝜀t, yt−j are independent for j ≥ 1. The stochastic process defined by Eq. (9) is
known as the GARCH process of order p, q and { is denoted}as GARCH (p, q). Let
Σt be a 𝜎-field associated with the sequence 𝜀t , 𝜀t−1 , … , , hence, 𝔼(𝜀2t |Σt ) = ht ,
i.e. conditional variance of ht is time varying rather than constant as volatility time
series studied in this work.
Given that linear GARCH models are not efficient in capturing possible asym-
metries in a time series, one alternative is to use non-linear GARCH techniques,
one of these is the Exponential GARCH approach, denoted as (EGARCH), which
is defined as follows:
p
� � � q
� � � �𝜀 � 2 𝜀t−i � � �
� t−i �
log ht = 𝛼0 + √ − + 𝜔 √ + 𝛽j log ht−j , (11)
i=1 ht−i 𝜋 ht−i j=1

where 𝛼i , 𝛽j , i = 1, 2, … , p, j = 1, 2, … , q and 𝜔 are positive constants. The Akaike

information criteria (AIC) and Bayesian information criteria (BIC) were used to find
the best model in terms of goodness of fit, as well as to add penalty criteria to the
number of selected parameters under minimization of these constants. For estima-
tion of the associated parameters, the maximum likelihood approach (MLE) was
used. Forecasts were carried out using one-step-ahead predictions for a fixed fore-
cast horizon.

3.4 Hybrid and wavelet models

3.4.1 Hybrid model

To improve predictions obtained by each predictive model individually, hybrid mod-

els are proposed, which exploit the advantages of each model and have been demon-
strated to be effective in different applications (Khashei & Bijari, 2011, 2012; Rubio
& Alba, 2022; Wang et al., 2013; Zhang, 2003). The first type of hybrid model used
in this work consider for example the use of ARIMA to extract linear components in
the time series and then forecast the associated residuals, being non-linear data with
heteroscedasticity, using for example the GARCH model. With this strategy, predic-
tion of linear and nonlinear patterns will be outperformed, thus obtaining a final
forecast with higher accuracy. Hybrid model Yt is then expressed as:
̃t + N
Yt = L ̃ t, (12)

where L ̃ t and Ñ t correspond to the linear and nonlinear components part of the
decomposition of the time series data. Let L ̃ t be the forecast obtained from ARIMA
in the volatility time series at t , hence, the corresponding residual 𝜀t are given by
̂t,
𝜀̂t = Yt − L (13)

13
Eurasian Economic Review (2023) 13:803–830 813

and are predicted applying the GARCH model and may be represented as
( )
𝜀̂t = ̂𝔣GARCH 𝜀̂t−1 , 𝜀̂t−2 , … , 𝜀̂t−n + Δt , (14)

where ̂𝔣GARCH is the nonlinear mapping corresponding to the GARCH model with
random error Δt and its forecast can be represented as N
̃ t . Accordingly, the hybrid
model (ARIMA-GARCH) is represented by
̂t = L
Y ̃t + N
̃ t, (15)

where L̃ t and N
̃ t represent predicted results of both linearity and nonlinearity based
on ARIMA and GARCH models, correspondingly. In the present work a second
alternative (GARCH-ARIMA) is considered. Similarly, to the definition of the
hybrid model presented above, but instead of applying ARIMA as a pre-processing
method before to apply GARCH model. Therefore, the most appropriate strategy for
predicting volatility using hybrid models is also analysed. In other words, the mod-
els associated with the terms L̃ t and N
̃ t are swapped with the GARCH and ARIMA
models, respectively.

3.4.2 Wavelet model

Another technique studied in this work, which also performs a mixed combination
of predictive models, is considered with the purpose of performing a comparative
study between these partitioned techniques. The technique propose below is related
̃ it , D
to the approach presented in Sect. 3.2. In this approach, each signal A ̃ it resulting
from the discrete wavelet decomposition (DWT) is predicted by applying ARIMA
or GARCH model to each component. These predictions are denoted by ̃ hw-ARIMA
t
or
̃
htw-GARCH , depending on the model used.
This work proposes to use ARIMA and GARCH models in a mixed strategy to
predict low and high frequency signals form (DWT). Considering each possible
combination in the selection of the models employed for each component, with the
objective of identifying which of these methodologies should be considered, for vol-
atility forecasting.
Formally, it is proposed to use models of the type:

̃
hw-ARIMA-GARCH ̃ ARIMA + D
=A ̃ GARCH + D
̃ GARCH ,
t 2t 2t 1t

where for each signal obtained from the wavelet decomposition, either high or low
frequency, the ARIMA and GARCH models are applied in a mixed form. In our
literature review, it was found that the usual way to apply these wavelet models is
by using an individual model, ARIMA or GARCH for low-frequency and high-
frequency components respectively, however, in this work a comparative study of
all possible options is performed. Selection of individual models for each wave is
described in Sects. 3.1 and 3.3.

13
814 Eurasian Economic Review (2023) 13:803–830

4 Stochastic volatility model

To make accuracy and performance comparisons between the proposed model and
one of the currently used for volatility forecasting, the stochastic volatility model
(SV) is introduced. Asset prices exhibit fluctuations in volatility over time, charac-
terized by periods of both high and low return variability. Stochastic volatility mod-
els capture this phenomenon by incorporating a latent volatility variable, which is
modelled as a stochastic process (Hoffman & Gelman, n.d.; Kim et al., 1998).
� �
h
𝜎 ∼ 𝜖t exp 2t ,
� �
ht+1 = 𝜇 + 𝜙� ht − 𝜇 �+ 𝛿t 𝜎,
h1 ∼ N 𝜇, √ 𝜎 2 ,
1−𝜙
𝜖t , 𝛿t ∼ N(0, 1),

where, 𝜖t represent the asset return white noise, at time t, 𝛿t is the shock on volatil-
ity and ht is the latent parameter for log volatility. The primary idea of this model
is identifying hidden factors that impact alterations in asset values. These undis-
closed elements, encompassing shifts in market sentiment, news, or other relevant
variables, play a role in determining the extent of market volatility. By representing
volatility as a dynamic and stochastic process, the SV model provides a more accu-
rate representation of financial markets, distinguishing it from simpler models that
assume a constant level of volatility.

5 Methodology

Prediction of volatility time series has been performed in different works using
machine learning and deep learning techniques as well as classical time series mod-
els. On the other hand, results obtained using a single forecasting strategy can be
improved by using mixed models, combining techniques that have different predic-
tive properties. Hybrid models have been used previously for time series forecast-
ing, usually by decomposing the original data into two components, associated with
the linear and nonlinear part of the series. In this manner, the use of two predictive
methods is combined, where usually the non-linear part is adjusted using models
able to capture strong fluctuations as ANNs for example (Zhang & Zhang, 2018).
Another hybrid technique found in the literature consider outputs from a GARCH
model as input of an ANN and conversely for forecasting volatility (Lu et al., 2016).
Stacking models have also been used to forecast volatility, where predictions are
dynamically selected from a set of trained machine learning models based on feature
extraction and selection (Aras, 2021).
Some authors proposed to use wavelet transform as a pre-processing to decom-
pose a time series into multiple high and low frequency waves, and depending on
the nature of the data different strategies are proposed to forecast each component.
For example, D. Liu et al. proposed to use SVM to forecast wind speed based just

13
Eurasian Economic Review (2023) 13:803–830 815

on the approximate signal. On the other hand, Glosten–Jagannathan–Runkle (GJR)-

GARCH model has been used together with wavelet transform, to predict finan-
cial returns, in this case each wavelet component is adjusted with the same model
(Berger, 2016). To forecast electricity prices, a wavelet transform based on ARIMA
have also been used, in this case the prediction is obtained applying ARIMA-
GARCH to the approximate signal and GARCH to the detailed signals (Tan et al.,
2010).
In this work, a comparative study between hybridization techniques based on
residuals decomposition and partitioned strategies according to wavelet transform
for volatility forecasting is presented. The strategy what indeed provide better pre-
dictions in terms of the accuracy metrics, MAPE, RMSE, and R2 is proposed. Ini-
tially it is proposed to use a technique based on wavelet transform and the ARIMA
and GARCH models is address, where ARIMA, being linear, is used to forecast the
low frequency wave, while GARCH is used for high frequency waves. For realized
volatility forecasting, the use of this strategy was not found in the literature found to
date, nor a detail description of the implementation process of this type of models in
the selected programming language Python. This choice of models has been found
to be the most natural and efficient for volatility forecasting, and consequently, it is
the one proposed in this work.
Most of the applications of the wavelet transform found in Python are mostly tar-
geted at digital image processing. Therefore, this paper describes how this imple-
mentation was done to serve also as a starting point for those interested in using
these types of partitioned models. In addition, given that in the related literature,
limited detail is provided regarding the functioning of the associated libraries for the
wavelet decomposition and recovery of the original signals. Figure 2 describes the
process of implementing the hybridization technique suggested.
One of the main challenges presented during the implementation of the wave-
let transform for time series forecasting is that there is limited development of the
wavelet transform when using the Python language. Most of the applications of the
wavelet transform found in Python are mostly targeted at digital image processing.
Therefore, this paper describes how this implementation was done to serve also as a
starting point for those interested in using these types of partitioned models. In addi-
tion, in the related literature, limited detail is provided regarding the functioning of
the associated libraries for the wavelet decomposition and recovery of the original
signals. Figure 2 describes the process of implementing the hybridization technique
proposed in this work. As can be seen in Fig. 2, to obtain the final prediction, first,
the respective decomposition of the daily returns is obtained by using the discrete
wavelet transform. Then, by using the discrete wavelet transform inverse, the origi-
nal signal is reconstructed through approximate and detailed components. Finally,
each signal is predicted by using, for example, ARIMA for approximate components
and GARCH for detailed components to get the prediction ̃ hw-ARIMA-GARCH
t
.

13
816 Eurasian Economic Review (2023) 13:803–830

Table 1  Descriptive parameters Closed price

for TSLA closed prices
Nrecords 3178
𝜇 59.95
𝜎 95.69
ymin 1.05
Q1 9.30
̃
x 16.36
Q3 27.01
ymax 409.97
Kurtosis 1.82
Skewness 1.79

6 Data

To carry out the empirical part of this investigation, the daily closing prices of
TSLA stocks were considered. Data was loaded directly from Yahoo Finance using
the free Python API, for a total of 3178 observations (between 2010-07-07 to 2023-
02-17). Descriptive parameters are calculated to obtain relevant information related
to measures of central tendency, dispersion, kurtosis and skewness of closed prices
(see Table 1). The values presented point to an asymmetrical (positive) data distri-
bution, with values of Q1 , Q2 and Q3 relatively close (in the first ten years considered
in the study, historical data show low values). Around 75% of the observations show
a closed price below USD 27, far from the maximum value observed (USD 409.97)
at the end of 2021.
Figure 3 shows for the TSLA closed price and trading volume strong fluctuations
associated with high volatility. Strong uptrend and a large trading volume ending

Fig. 3  Time series for TSLA closed price (left) and trading volume (right)

13
Eurasian Economic Review (2023) 13:803–830 817

Table 2  Normality tests (Jarque–Bera test and skewness and kurtosis tests), stationarity/existence of unit
root (ADF test and KPSS test) and independence test (BDS test) for TSLA closed prices
Test Normality Unit root/stationary Independence
Kurtosis skewness Jarque–Bera ADF KPSS BDS (Dim.2–Dim.6)

Statistic 11.62 28.80 2139.65 − 0.98 6.15 11.67–23.88

p-value 0.00* 0.00* 0.00* 0.75 – 0.00*

*Rejected for the significance levels 1%, 5%, 10%

around 2021 can be identified. In terms of trends, we can identify three clearly dis-
tinct moments. Until 2020, historical data shows a relatively constant trend (with
low closed price). Globally, there is a marked growth between 2020 and the end of
2021, although with some fluctuations, namely in early 2021 (due to the instability
of the financial markets, partly caused by the COVID-19 pandemic). From the end
of 2021, the data is marked by an abrupt drop in the closed price.
To analyse some features of the closed price time series, Table 2 contains the
statistic test and the following hypothesis tests: Normality tests (Jarque–Bera test
and Skewness and Kurtosis tests), Stationarity/Existence of unit root (ADF test and
KPSS test) and independence test (BDS test). As expected, for any significance
level, the normality, stationarity, and independence tests are rejected. In fact, there
is statistical evidence to: (i) do not reject the nonnormality of the distribution of
the data (with rejection of the null hypothesis, which indicates normal behaviour, in
all tests performed); (ii) assume the non-stationarity (due to the null hypothesis not
being rejected when doing ADF test, and due to the statistical value corresponding
to KPSS being superior to the critical reference values); (iii) infer about the non-iid,
since the null hypothesis of the data being iid has been rejected through BDS test.
According to (Karasan, n.d., 2005), to model volatility we need to calculate the
return volatility, which is also known as realized volatility. Realized volatility is the
square root of realized variance, which is the sum of squared return. Realized vola-
tility is used to calculate the performance of the volatility prediction method, it is
denoted by ht in the model description of GARCH (Eq. 9).

7 Numerical results

To continue the study, based on historical data on TSLA closing prices, described
in previous section, the time series of the daily returns and realized volatility
were obtained. Figure 4 shows daily returns plot, whose values are around the
zero mean. As can be seen, the variance is a function of time, which confirms the
effect of heteroscedasticity in the time series, and the associated histogram shows
that the daily returns have a non-normal distribution. Regarding realized volatility
time series, note the peaks of values observed during the period 2020/2021 (insta-
bility of the financial markets, partly motivated by the COVID-19 pandemic).

13
818 Eurasian Economic Review (2023) 13:803–830

Fig. 4  Daily return series (left) and KDE histogram (right) plots for TSLA stock

Fig. 5  Realized volatility time series for TSLA stock. Graphical representation of annual box plots

Table 3  Descriptive parameters Daily returns Realized volatility

for TSLA daily returns and
realized volatility Nrecords 3181 3177
𝜇 0.21 3.06
𝜎 3.61 1.90
ymin − 21.06 0.33
Q1 − 1.55 1.76
̃
x 0.12 2.57
Q3 1.93 3.83
ymax 24.39 14.78
Kurtosis 5.01 5.37
Skewness 0.32 1.87

13
Eurasian Economic Review (2023) 13:803–830 819

Table 4  Normality tests (Jarque–Bera test and skewness and kurtosis tests), stationarity/existence of unit
root (ADF test and KPSS test) and independence test (BDS test) for TSLA daily returns and realized
volatility
Test Normality Unit root/station- Independence
ary
Kurtosis Skewness Jarque–Bera ADF KPSS BDS (Dim.2–Dim.6)

D. return statistic 18.86 8.28 3385.62 − 56.93 0.08 0.44–1.36

p− value 0.00* 0.00* 0.00* 0.00* – 0.65–0.17
R. volatility Statistic 19.38 29.58 5681.73 − 6.10 0.10 4.84–7.47
p-value 0.00* 0.00* 0.00* 0.00* – 0.00*

* Means rejected for the significance levels 1%, 5%, 10%

This detail can be better perceived when evaluating the annual box plots
(Fig. 5), where a considerable sample amplitude is observed in 2020 (with several
outliers observations above the maximum barrier of the box-plot) and a notable
interquartile amplitude, in comparison with the other years.
Complementing the graphical representation, some descriptive statistical
measures are presented in Table 3. Based on these values, a positive kurtosis (lep-
tokurtic distribution) and a slight asymmetry to the right are observed.
As was done in previous section, to analyse some features of the time series,
Table 4 contains the statistic test and the following hypothesis tests: Normal-
ity tests (Jarque–Bera test and Skewness and Kurtosis tests), Stationarity/Exist-
ence of unit root (ADF test and KPSS test) and independence test (BDS test). As
expected, for any significance level, the normality is rejected for the two series
under study, concluding that it does not have a normal distribution.
In addition, for both time series, ADF test and KPPS tests confirm that the series
is stationary (due to the null hypothesis being rejected when doing ADF test, and
due to the statistical value corresponding to KPSS being less to the critical refer-
ence values). When testing for Independency, for daily returns it is concluded that
the null hypothesis is not rejected for any significance level and therefore the data
are i.i.d. For realized volatility time series, there is statistical evidence to infer about

Table 5  Normality tests on daily returns and realized volatility for TLSA stocks
Test Shapiro–Wilk D’Agostino’s K2 Anderson–Darling
Significance Critical value
level

p-value 0.00000 0.00000 42.514 15 0.575

10 0.655
5 0.786
2.5 0.917
1 1.091

p-values for Shapiro–Wilk, D’Agostino’s K2 and Anderson–Darling tests are registered

13
820 Eurasian Economic Review (2023) 13:803–830

the non-iid, since the null hypothesis of the data being iid has been rejected through
BDS test. Different normality test such as: Shapiro–Wilk, D’Agostino’s K ­ 2 and
Anderson–Darling tests, were implemented to verify normality and results are pre-
sented in Table 5. As can be seen, based on the null hypothesis:
H0 ∶ dataset has normal distribution

Shapiro–Wilk, and D’Agostino’s K 2 test yields p− values far below the signifi-
cance threshold of 0.05. Therefore, the null hypothesis is rejected, i.e. the distri-
bution of daily returns is not normal. On the other hand, it can be seen that daily
returns distributions are peaked around the mean, which is confirmed with positive
kurtosis values larger than 2, it means most of the daily return values are concen-
trated in the mean. Furthermore, this implies a high level of risk but the possibil-
ity of higher returns due to large price movements. Since the statistic (42.514), on
Anderson–Darling test, is greater than all critical values for different significance
levels, non-normality is confirmed.

7.1 Application of ARIMA model

Daily returns and realized volatility for TLSA show stationary signals, with p-values
less than 0.05 on Augmented Dickey-Fuller test. Therefore, the null hypothesis
H0: dataset has nonstationary distribution
is rejected, thus each share shows a stationary signal; consequently, an integration
order greater than zero is not necessary. The ARIMA model was trained based on
the Box and Jenkins strategy to get the best model in terms of goodness of fit. Opti-
mal orders p, d, q were obtained based on the Akaike information criterion (AIC),
which provides best orders for model quality, based on the minimization of the AIC.

Fig. 6  Real vs. ARIMA adjustment for TSLA realized volatility (left). Correlation plots for the test set
and prediction with its corresponding R2 value (right)

13
Eurasian Economic Review (2023) 13:803–830 821

Table 6  Accuracy metrics MAPE, MAE, RMSE, and R2 for TSLA realized volatility forecast using
ARIMA and GARCH
Model MAPE MAE RMSE R2

ARIMA 0.25 0.75 1.03 0.71

GARCH 0.25 0.99 1.24 0.65

Fig. 7  Real vs. GARCH adjustment for TSLA realized volatility (left). Real time series is represented
in green and its prediction in blue colour. Correlation plots for the test set and prediction with its cor-
responding R2 value (right)

The minimum AIC was obtained by doing a loop over a fixed range for p, q where
the order d was set to zero since no differentiation was necessary given that Dickey-
Fuller test confirmed stationarity.
ARIMA was adjusted using the TSLA stock volatility training set composed of
all the available historical data, except for the last 21 days considered for the test
set or prediction horizon. One-step-ahead forecasting was considered in this work.
The Best ARIMA model obtained was ARIMA(4, 0, 4) for realized volatility predic-
tion. Figure 6 shows realized volatility prediction and correlation plots between the
original and predicted time series. Goodness of fit and simulation performance can
be confirmed with the R2 and accuracy metrics, such as MAPE, MAE, and RMSE
(Table 6). As can be seen, ARIMA shows good performance forecasting implicit lin-
ear patterns in TSLA realized volatility, maintaining trends for a long-term horizon;
nevertheless, it is essential to use another method to detect non-linear patterns.

7.2 Application of GARCH models

Similar to ARIMA training and considering the same prediction horizon, the best
orders p, q for the GARCH model were obtained by minimization of the AIC

13
822 Eurasian Economic Review (2023) 13:803–830

Fig. 8  Real vs. ARIMA, and GARCH adjustment for TSLA realized volatility (left). Correlation plots
with its corresponding R2 value (right)

coefficient. Parameters obtained for the ARCH model with a prediction horizon of
21 days were p = 4, q = 2, considering a zero-volatility process and the exponential
version of the GARCH model as optimal parameters after a grid search was applied.
Coefficients for the model were obtained based on the maximum likelihood estima-
tion (MLE). Figure 7 shows model adjustment for the real time series in the test
set and correlation plots between the original and predicted realized volatility. It is
remarkable the difference between goodness of fit between GARCH and ARIMA
models, although in terms of error metrics they are only slightly different (Table 6).
ARIMA remains as the model with the best goodness-of-fit. Figure 8 shows ARIMA
and GARCH model predictions and their correlations with the original time series.

7.3 Application of hybrid and partitioned models

A first approximation to combine ARIMA and GARCH models is to predict real-

ized volatility first using for example ARIMA and then predict its residuals with
ARCH to have the first hybrid model implemented in this work denoted by ARIMA-
GARCH (see Sect. 3.4). The second hybrid model can be obtained, changing roles
for the first and second model applied to the original time series and its forecasted
residuals, this model will be denoted by GARCH-ARIMA. This second option has

Table 7  Accuracy metrics Model MAPE MAE RMSE R2

MAPE, MAE, RMSE, and R2
for realized volatility forecast
ARIMA 0.25 0.75 1.03 0.71
using individual and hybrid
models GARCH 0.25 0.99 1.24 0.65
ARIMA-GARCH 0.24 1.07 1.26 0.70
GARCH-ARIMA 0.25 0.99 1.24 0.65

13
Eurasian Economic Review (2023) 13:803–830 823

the disadvantage that residuals are extremely small, and therefore the contribution
of ARIMA is not significant, resulting in the same accuracy results provided by
GARCH (see Table 7).
The second option for combining the econometric models proposed in this
work is to consider the wavelet transform as a pre-processing method for the time
series of interest. Daily return time series is decomposed in an approximate signal
named At and a detailed signal denoted by Dt using the wavelet transform. Usu-
ally, detailed signal has higher frequency and therefore in this work it is proposed
to use GARCH model to predict this component and ARIMA as a linear model for
the low frequency signal. This first approximation will be denoted by W-ARIMA-
GARCH. It means, over the signals At and Dt , ARIMA and GARCH models are
applied respectively. By W-GARCH-ARIMA shall be denoted the opposite, for
selecting predictive models for low and high frequency decomposed signals (see
Fig. 2). In addition, it was investigated the case where for each of the wavelets
obtained by wavelet transform the same model, either ARIMA or GARCH, was
applied to study the behaviour of these predictions. Prediction results are pre-
sented in Fig. 9, where the original TSLA time series is plotted over the predic-
tion and variance explained by each model is at the right side.
For W-ARIMA model, the following parameters were obtained from the train-
ing datasets for TSLA daily returns after a manual grid search in pursuit of the
best decomposition and models: Daubechy wave db2 cast the orders (4, 1, 4) and
(4, 0, 4) for high and low frequency waves. Figure 10 shows time series decom-
position using the db2-Daubechy filter. Next, the W-GARCH model was imple-
mented, and the parameters obtained were: Daubechy wave db3, (4, 4) and (4,
4) for low and high frequency signals considering zero mean model and expo-
nential GARCH. Figure 11 shows the model adjustment for the real time series,
with its respective correlation plots. As can be seen, W-GARCH prediction are
strongly penalizing the high values of the time series. A possible reason can be

Fig. 9  Real vs. ARIMA-GARCH and GARCH-ARIMA predictions for TSLA realized volatility (left).
Correlation plot with its corresponding R2 value (right)

13
824 Eurasian Economic Review (2023) 13:803–830

Fig. 10  Decomposition with db2-Daubechy filter and ARIMA-WAVELET adjustment for each decom-
position

Fig. 11  Real vs. W-GARCH and W-ARIMA models predictions for TSLA realized volatility. Correlation
plots for the test set and prediction with its corresponding R2 value (right)

to use GARCH process to adjust low frequency signal from the decomposition.
Comparing with W-ARIMA an improvement can be seen and confirmed with
the R2 as goodness of fit measure and accuracy metrics (see Table 8). Therefore,

13
Eurasian Economic Review (2023) 13:803–830 825

Table 8  Accuracy metrics MAPE, MAE, RMSE, and R2 for realized volatility forecast using the
W-ARIMA, W-GARCH models
Model MAPE MAE RMSE R2

W-ARIMA 0.26 0.79 1.00 0.68

W-GARCH 0.35 1.80 2.38 0.36

Fig. 12  Real vs. W-GARCH-ARIMA and W-ARIMA-GARCH models predictions for TSLA realized
volatility. Correlation plots with its corresponding R2 value (right)

Table 9  Accuracy metrics Model MAPE MAE RMSE R2

MAPE, MAE, RMSE, and R2
for realized volatility forecast
ARIMA 0.25 0.75 1.03 0.71
using all the models studied
GARCH 0.25 0.99 1.24 0.65
SV 0.44 1.12 1.73 0.24
W-ARIMA 0.26 0.79 1.00 0.68
W-GARCH 0.35 1.80 2.38 0.36
ARIMA-GARCH 0.24 1.07 1.26 0.70
GARCH-ARIMA 0.25 0.99 1.24 0.65
W-ARIMA-GARCH 0.24 1.27 1.59 0.76
W-GARCH-ARIMA 0.37 1.07 1.34 0.62
W-ARIMA-SV 0.37 2.53 3.02 0.13

applying ARIMA to each component in terms of goodness of fit is better than

using GARCH on each one.
For W-ARIMA-GARCH model the following were parameters obtained for
TSLA daily returns: Daubechy wave db3, orders (4, 2), with optional exogenous

13
826 Eurasian Economic Review (2023) 13:803–830

regressors and exponential GARCH) version for high frequency signal and (4, 0,
4) as ARIMA orders for low frequency. W-GARCH-ARIMA model yielded the
following parameters for the best model: Daubechy wave db2, (4, 1, 4) as ARIMA
orders for approximate signal and (4, 2), zero mean, and exponential GARCH for
the low frequency component.
Figure 12 shows models adjustment for the real time series of realized volatil-
ity and the respective correlation plots. Goodness of fit can be confirmed with R2
and accuracy metrics MAPE, MAE, and RMSE are resumed in Table 9 for all the
models studied in this work. Thus, applying ARCH and ARIMA process in high
and low frequency signal provides better predictions in terms of goodness-of-fit.
As before, applying a GARCH process on the low frequency decomposition fails
to give optimal prediction results.
Finally, the stochastic volatility model was implemented for the same prediction
horizon, using rolling-forecast. However, the model proposed in this work showed
better error metrics and goodness of fit. For the implementation of the SV model,
defined in Sect. 4, 𝜙 ∼ uniform(−1, 1), 𝜎 ∼ cauchy(0, 5) and 𝜇 ∼ cauchy(0, 10) were
considered (Fig. 13).

8 Conclusions

The present work provides a starting point for any research related to the use of
hybrid and wavelet transform models for volatility forecasting. It also proposed
ARIMA and GARCH models which do not require significant computational train-
ing costs as alternative strategies to predict each signal. This same strategy can be
used considering even any other pair of models, e.g. machine learning or deep learn-
ing. A detailed description has been made with outlines, which show how the imple-
mentation of these models in Python language should be done due to the fact that
there is very little in the literature on the subject.

Fig. 13  Real vs. SV model with 𝜙 ∼ uniform(−1, 1), 𝜎 ∼ cauchy(0, 5) and 𝜇 ∼ cauchy(0, 10)

13
Eurasian Economic Review (2023) 13:803–830 827

ARIMA models, when used in combination with the wavelet transform to pre-
dict realized volatility, will provide better predictions than those obtained with the
GARCH model in terms of goodness of fit and each accuracy metric. By using mul-
tiple hybridization techniques between the proposed models, prediction scores can
be improved. Selection of models to be used in each component of the wavelet trans-
form decomposition is crucial to obtain more accurate predictions. In conclusion,
the best alternative is to forecast low frequency signals with linear models such as
ARIMA and high frequency signals with models that can capture strong nonlinear
patterns, such as the GARCH model.
Models based on wavelet transform are compared with hybrid models, based on
forecasting the original time series using for example a linear model and then its
residual using a model able to detect strong fluctuations, such as GARCH. The two
possible combinations were studied, to reach the conclusion that in order to obtain
improved predictions, the model that should always be applied first is the linear
model, in this case ARIMA. The hybrid model was compared with each model indi-
vidually and the partitioned wavelet transform approach. In terms of goodness of fit,
the best model for forecasting volatility is the wavelet transform where the approxi-
mate signal is predicted with a linear model, in this case ARIMA and the detailed
high frequency using GARCH.
The present W-ARIMA-GARCH model has been applied to Tesla stock volatility,
to quantify risk, however, it can be applied to any other time series with strong fluc-
tuations, such as cryptocurrencies, for example. One of the advantages of this model
is that it takes advantage of the ARIMA and GARCH models to capture different
types of patterns in the time series. In addition, another advantage of implementing
this type of forecasts is that they can be optimized, allowing to obtain predictions in
short CPU time. For example, for algorithmic trading firms, where fast predictions
must be executed, to reduce risk and detect market and arbitrage opportunities, it is
of high importance.
As future work, it is proposed to make use of this technique in different applica-
tions where it is not yet used, to prove its effectiveness. Investigate which alterna-
tive models can be used in this type of decomposition. Also, study techniques based
on genetic algorithms and stacked neural networks to compare their performance in
comparison with the models studied in this work. The use of sentiment analysis can
be used to refine a financial decision, therefore as future work is also proposed to
study natural language processing and sentiment analysis to obtain even more accu-
rate predictions.
Acknowledgements This work is partially financed by national funds through FCT—Fundação para a
Ciência e a Tecnologia under the project UIDB/00006/2020. L.R. is supported by Universidad del Norte.
Research Agenda 2021-009. The paper has also benefited from discussions at the 42nd EBES Confer-
ence, in Lisbon.

Funding Open access funding provided by FCT|FCCN (b-on). This work is partially financed by national
funds through FCT—Fundação para a Ciência e a Tecnologia under the project UIDB/00006/2020.

Data availability Datasets used in this work were acquired through the use of the free Yahoo Finance
Python API.

13
828 Eurasian Economic Review (2023) 13:803–830

Declarations
Conflict of interest The authors declare no conflict of interest.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as
you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com-
mons licence, and indicate if changes were made. The images or other third party material in this article
are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is
not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission
directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​
ses/​by/4.​0/.

References
Alshammari, T. S., Ismail, M. T., Al-wadi, S., Saleh, M. H., & Jaber, J. J. (2020). Modeling and forecast-
ing Saudi stock market volatility using wavelet methods. The Journal of Asian Finance, Economics
and Business, 7(11), 83–93. https://​doi.​org/​10.​13106/​jafeb.​2020.​vol7.​no11.​083
Al Wadi, S., Ismail, M. T., Altaher, A. M., & Addul Karim, S. A. (2010). Forecasting volatility data
based on Wavelet transforms and ARIMA model. In 2010 International Conference on Science and
Social Research (CSSR 2010) (pp. 86–90). https://​doi.​org/​10.​1109/​CSSR.​2010.​57739​09
Aras, S. (2021). Stacking hybrid GARCH models for forecasting Bitcoin volatility. Expert Systems with
Applications, 174, 114747. https://​doi.​org/​10.​1016/J.​ESWA.​2021.​114747
Asteriou, D., & Hall, S. G. (2016). ARIMA models and the Box–Jenkins methodology. In Applied Econo-
metrics (pp. 275–296). Macmillan Education UK. https://​doi.​org/​10.​1057/​978-1-​137-​41547-9_​13
Bathla, G. (2020). Stock price prediction using LSTM and SVR. In 2020 Sixth International Confer-
ence on Parallel, Distributed and Grid Computing (PDGC) (pp. 211–214). https://​doi.​org/​10.​1109/​
PDGC5​0313.​2020.​93158​00
Berger, T. (2016). Forecasting based on decomposed financial return series: A wavelet analysis. Journal
of Forecasting, 35(5), 419–433. https://​doi.​org/​10.​1002/​for.​2384
Box, G. E. P., Jenkins, G. M., & Reinsel, G. C. (2008). Time Series Analysis: Forecasting and Control.
Wiley.
Chniti, G., Bakir, H., & Zaher, H. (2017). E-commerce time series forecasting using LSTM neural net-
work and support vector regression. In Proceedings of the international conference on big data and
internet of thing (pp. 80–84). https://​doi.​org/​10.​1145/​31756​84.​31756​95
Christensen, K., Siggaard, M., & Veliyev, B. (2021). A machine learning approach to volatility forecast-
ing. Journal of Financial Econometrics. https://​doi.​org/​10.​1093/​jjfin​ec/​nbac0​20
Daubechies, I. (1992). Ten lectures on wavelets. Society for Industrial and Applied Mathematics. https://​
doi.​org/​10.​1137/1.​97816​11970​104
de Oliveira, E. M., & Cyrino Oliveira, F. L. (2018). Forecasting mid-long term electric energy consump-
tion through bagging ARIMA and exponential smoothing methods. Energy, 144, 776–788. https://​
doi.​org/​10.​1016/J.​ENERGY.​2017.​12.​049
Engle, R. F. (1982). Autoregressive conditional heteroscedasticity with estimates of the variance of
United Kingdom inflation. Econometrica, 50(4), 987. https://​doi.​org/​10.​2307/​19127​73
Gaoyu, Z., Qiongfei, L., Qing, L., & Zhizhao, Z. (2009). High frequency financial time series forecast-
ing via particle filtering. In 2009 International Conference on Information Management, Innovation
Management and Industrial Engineering (pp. 62–65). https://​doi.​org/​10.​1109/​ICIII.​2009.​477
Gavrishchaka, V. V., & Banerjee, S. (2006). Support vector machine as an efficient framework for stock
market volatility forecasting. Computational Management Science, 3(2), 147–160. https://​doi.​org/​
10.​1007/​s10287-​005-​0005-5
Guo, J., Xie, Z., Qin, Y., Jia, L., & Wang, Y. (2019). Short-term abnormal passenger flow prediction
based on the fusion of SVR and LSTM. IEEE Access, 7, 42946–42955. https://​doi.​org/​10.​1109/​
ACCESS.​2019.​29077​39

13
Eurasian Economic Review (2023) 13:803–830 829

Hafner, C. M. (2009). Causality and forecasting in temporally aggregated multivariate GARCH pro-
cesses. Econometrics Journal, 12(1), 127–146. https://​doi.​org/​10.​1111/j.​1368-​423X.​2008.​00276.x
Hoffman, M. D., & Gelman, A. (n.d.). The no-U-turn sampler: adaptively setting path lengths in Hamil-
tonian Monte Carlo. http://​mcmc-​jags.​sourc​eforge.​net
Hyup Roh, T. (2007). Forecasting the volatility of stock price index. Expert Systems with Applications,
33(4), 916–922. https://​doi.​org/​10.​1016/J.​ESWA.​2006.​08.​001
Kang, S. H., Kang, S. M., & Yoon, S. M. (2009). Forecasting volatility of crude oil markets. Energy Eco-
nomics, 31(1), 119–125. https://​doi.​org/​10.​1016/J.​ENECO.​2008.​09.​006
Karasan, A. (n.d.). Machine learning for financial risk management with Python: algorithms for modeling
risk.
Khashei, M., & Bijari, M. (2011). A novel hybridization of artificial neural networks and ARIMA models
for time series forecasting. Applied Soft Computing, 11(2), 2664–2675. https://​doi.​org/​10.​1016/J.​
ASOC.​2010.​10.​015
Khashei, M., & Bijari, M. (2012). A new class of hybrid models for time series forecasting. Expert Sys-
tems with Applications, 39(4), 4344–4357. https://​doi.​org/​10.​1016/J.​ESWA.​2011.​09.​157
Kim, S., Shepherd, N., & Chib, S. (1998). Stochastic volatility: Likelihood inference and comparison
with ARCH models. Review of Economic Studies, 65(3), 361–393. https://​doi.​org/​10.​1111/​1467-​
937X.​00050
Li, X., & Tang, P. (2020). Stock index prediction based on wavelet transform and FCD-MLGRU. Journal
of Forecasting, 39(8), 1229–1237. https://​doi.​org/​10.​1002/​for.​2682
Liu, Y. (2019). Novel volatility forecasting using deep learning–Long short term memory recurrent neural net-
works. Expert Systems with Applications, 132, 99–109. https://​doi.​org/​10.​1016/J.​ESWA.​2019.​04.​038
Lu, X., Que, D., & Cao, G. (2016). Volatility forecast based on the hybrid artificial neural network and
GARCH-type models. Procedia Computer Science, 91, 1044–1049. https://​doi.​org/​10.​1016/J.​
PROCS.​2016.​07.​145
Luzia, R., Rubio, L., & Velasquez, C. E. (2023). Sensitivity analysis for forecasting Brazilian electricity
demand using artificial neural networks and hybrid models based on autoregressive integrated mov-
ing average. Energy, 274, 127365. https://​doi.​org/​10.​1016/J.​ENERGY.​2023.​127365
Mallat, S. G. (n.d.). Fundamental papers in wavelet theory.
Mallat, S. G., & Peyré, G. (n.d.). A wavelet tour of signal processing: the sparse way.
Maria, F. C., & Eva, D. (n.d.). Exchange-rates forecasting: exponential smoothing techniques and
ARIMA models.
Montgomery, D. C., Jennings, C. L., & Kulahci, M. (n.d.). Wiley series in probability and statistics.
Qi, C., Ren, J., & Su, J. (2023). GRU neural network based on CEEMDAN-wavelet for stock price pre-
diction. Applied Sciences, 13(12), 7104. https://​doi.​org/​10.​3390/​app13​127104
Ramos-Pérez, E., Alonso-González, P. J., & Núñez-Velázquez, J. J. (2019). Forecasting volatility with a
stacked model based on a hybridized artificial neural network. Expert Systems with Applications,
129, 1–9. https://​doi.​org/​10.​1016/J.​ESWA.​2019.​03.​046
Rubio, L., & Alba, K. (2022). Forecasting selected colombian shares using a hybrid ARIMA-SVR model.
Mathematics, 10(13), 2181. https://​doi.​org/​10.​3390/​math1​01321​81
Rubio, L., Gutiérrez-Rodríguez, A. J., & Forero, M. G. (2021). EBITDA index prediction using exponen-
tial smoothing and ARIMA model. Mathematics, 9(20), 2538. https://​doi.​org/​10.​3390/​math9​202538
Tan, Z., Zhang, J., Wang, J., & Xu, J. (2010). Day-ahead electricity price forecasting using wavelet trans-
form combined with ARIMA and GARCH models. Applied Energy, 87(11), 3606–3610. https://​doi.​
org/​10.​1016/J.​APENE​RGY.​2010.​05.​012
Tang, L.-B., Tang, L.-X., & Sheng, H.-Y. (2009). Forecasting volatility based on wavelet support vector
machine. Expert Systems with Applications, 36(2), 2901–2909. https://​doi.​org/​10.​1016/j.​eswa.​2008.​
01.​047
Wang, L., Zou, H., Su, J., Li, L., & Chaudhry, S. (2013). An ARIMA-ANN hybrid model for time series
forecasting. Systems Research and Behavioral Science, 30(3), 244–259. https://​doi.​org/​10.​1002/​sres.​2179
Wang, Y., & Guo, Y. (2020). Forecasting method of stock market volatility in time series data based on
mixed model of ARIMA and XGBoost. China Communications, 17(3), 205–221. https://​doi.​org/​10.​
23919/​JCC.​2020.​03.​017
Xiao, Y., Xiao, J., Liu, J., & Wang, S. (2014). A multiscale modeling approach incorporating ARIMA
and ANNS for financial market volatility forecasting. Journal of Systems Science and Complexity,
27(1), 225–236. https://​doi.​org/​10.​1007/​s11424-​014-​3305-4

13
830 Eurasian Economic Review (2023) 13:803–830

Yu, H., Ming, L. J., Sumei, R., & Shuping, Z. (2020). A hybrid model for financial time series fore-
casting—Integration of EWT, ARIMA with the improved ABC optimized ELM. IEEE Access, 8,
84501–84518. https://​doi.​org/​10.​1109/​ACCESS.​2020.​29875​47
Zhang, P. G. (2003). Time series forecasting using a hybrid ARIMA and neural network model. Neuro-
computing, 50, 159–175. https://​doi.​org/​10.​1016/​S0925-​2312(01)​00702-0
Zhang, Y., & Zhang, J. (2018). Volatility forecasting of crude oil market: A new hybrid method. Journal
of Forecasting, 37(8), 781–789. https://​doi.​org/​10.​1002/​for.​2502

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

13