Annals of Operations Research (2021) 306:295–320

https://doi.org/10.1007/s10479-021-04000-8

S.I. : REGRESSION METHODS BASED ON OR TECHNIQUES

A differential evolution‑based regression framework

for forecasting Bitcoin price

R. K. Jana1 · Indranil Ghosh2 · Debojyoti Das3

Accepted: 16 February 2021 / Published online: 1 March 2021

© The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature 2021

Abstract
This research proposes a differential evolution-based regression framework for forecasting
one day ahead price of Bitcoin. The maximal overlap discrete wavelet transformation first
decomposes the original series into granular linear and nonlinear components. We then
fit polynomial regression with interaction (PRI) and support vector regression (SVR) on
linear and nonlinear components and obtain component-wise projections. The sum of these
projections constitutes the final forecast. For accurate predictions, the PRI coefficients
and tuning of the hyperparameters of SVR must be precisely estimated. Differential evo-
lution, a metaheuristic optimization technique, helps to achieve these goals. We compare
the forecast accuracy of the proposed regression framework with six advanced predictive
modeling algorithms- multilayer perceptron neural network, random forest, adaptive neu-
ral fuzzy inference system, standalone SVR, multiple adaptive regression spline, and least
absolute shrinkage and selection operator. Finally, we perform the numerical experimenta-
tion based on—(1) the daily closing prices of Bitcoin for January 10, 2013, to February 23,
2019, and (2) randomly generated surrogate time series through Monte Carlo analysis. The
forecast accuracy of the proposed framework is higher than the other predictive modeling
algorithms.

Keywords Differential evolution · Polynomial regression with interaction · Support vector

regression · Maximal overlap discrete wavelet transformation · Bitcoin

* R. K. Jana
rkjana1@gmail.com
Indranil Ghosh
fri.indra@gmail.com
Debojyoti Das
debojyoti.das@iimb.ac.in
1
Operations and Quantitative Methods Area, Indian Institute of Management Raipur, Atal Nagar,
Raipur, CG 493661, India
2
Department of Operations Management and IT, Calcutta Business School, Kolkata, WB 743503,
India
3
Finance and Accounting Area, Indian Institute of Management Bangalore, Bengaluru,
Karnataka 560076, India

13
Vol.:(0123456789)
296 Annals of Operations Research (2021) 306:295–320

1 Introduction

Cryptocurrencies are well known for their hedging, portfolio diversification, and risk miti-
gation capabilities across the literature (Bouoiyour et al. 2019; Bouri et al. 2018; Koutmos
2019). Bitcoin, the most popular cryptocurrency, has drawn crucial market players’ atten-
tion at different hierarchies for accomplishing these activities. Simultaneously, the litera-
ture predominantly reports Bitcoin’s return behavior as chaotic and multifractal (Al-Yahaee
et al. 2018; Filho et al. 2018) and more volatile than other financial assets (Chaim and
Laurini 2018).
Several researchers report the critical properties of Bitcoin and its interplay with other
assets (Tiwari et al. 2018; Wang et al. 2019; Horra et al. 2019; Poyser 2019; Zhang et al.
2020). The assessment of the price dynamics of Bitcoin remains very challenging due to
the high degree of nonstationary, nonlinear, volatile behaviors and the influence of various
uncontrollable factors (Matta et al. 2016; Cretarola and Figà-Talamanca 2019). The con-
ventional predictive models like moving average (MA), Autoregressive (AR), autoregres-
sive integrated moving average (ARIMA) fail to produce accurate forecasts (Ghosh et al.
2019). Besides, these models suffer from several shortcomings like the strict requirement
of the stationary of datasets, homoscedasticity of errors, etc. (Filho et al. 2018). Research-
ers propose machine learning, computational intelligence, deep learning, and hybrid mod-
els (Ghosh et al. 2017, 2019; Kumar and Ravi 2017) to alleviate these limitations.
This work presents an efficient regression-based predictive modeling framework for
forecasting a one-day ahead price of Bitcoin. An accurate estimation of a one-day ahead
price forecast may help the daily traders make better investment decisions and more profit.
Support vector regression (SVR) and polynomial regression with interactions (PRI) help to
learn the inherent governing patterns on the decomposed time series. The maximal over-
lapping discrete wavelet transformation (MODWT) decomposes the original time series.
We use polynomial regression with interactions on linear sub-components and SVR on
nonlinear sub-components to generate forecasts. Tuning of hyperparameters of SVR is nec-
essary to ensure superior forecast accuracy from the two regression models. Also, we need
to estimate the PRI coefficients. DE helps us to achieve these two tasks. We combine the
two regression models’ forecasts to generate the final prediction and carry out the proposed
approach’s comparative performance evaluation with MLP, RF, ANFIS, SVR, MARS, and
LASSO, considering the original and five distinct surrogate time series simulated from the
original Bitcoin price data.
This research’s significant contribution is the proposition of a novel predictive mod-
eling approach by combining the optimization and regression frameworks in a granular
set up for time series forecasting. The proposed approach consists of four distinct compo-
nents—MODWT for granular decomposition, SVR and PRI for learning pattern, and DE
for hyperparameter tuning of SVR and the PRI model’s estimation. It is different from the
existing wavelet decomposition-based approaches in terms of using two distinct predictive
modeling algorithms on appropriate granular components (Ghosh et al. 2017a, 2019). This
approach is useful for predictive modeling of time series displaying an excess degree of
volatility like the Bitcoin for making a more accurate forecast.
The rest of the paper’s organization is as follows: Sect. 2 presents the previous research.
Section 3 describes data and critical statistical properties of Bitcoin prices. Section 4 pre-
sents the research methodology, while Sect. 5 reports the results obtained from the pro-
posed framework. Section 6 concludes the paper.

13
Annals of Operations Research (2021) 306:295–320 297

2 Previous research

This section clusters the pertinent work to inspect Bitcoin’s temporal characteristics,
interaction dynamics with other assets, and predictive modeling. Several research-
ers study the temporal characteristics of Bitcoin (Bariviera 2017; Nadarajah and Chu
2017; Takaishi 2018). A section of literature proves that the Bitcoin market is ineffi-
cient (Urquhart 2016; Kristoufek et al. 2018; Tiwari et al. 2018). Researchers examine
the same market’s multifractal nature through multifractal detrended fluctuation analy-
sis (Filho et al. 2018; Kristjanpoller et al. 2020; Takaishi 2018). The findings suggest
that the degree of multifractal nature is more substantial in Bitcoin than several other
stock market indices. A copula-based Granger causality test explores the causal inter-
play between Bitcoin’s attention manifested through google trends and Bitcoin returns
from January 1, 2013, to December to December 31, 2017 (Dastgir et al. 2019). The
study reveals the existence of bidirectional causal nexus in the left and right tail of the
distribution.
The interaction between Bitcoin returns and economic policy uncertainty (EPU)
explores the hedging capacity of Bitcoin (Demir et al. 2018; Paule Vianez et al. 2020).
Wang et al. (2019) measure the risk spillover effect from EPU to Bitcoin and selected
equity market uncertainty index, US EPU index, and implied volatility index as proxies
for EPU. The results indicate the presence of negligible spillover from EPU to Bitcoin.
So, Bitcoin may act as a safe asset for investment under EPU shocks. The Bayesian
structural time series approach inspects the temporal dynamics and dependence struc-
ture of Bitcoin prices on other assets (Poyser 2019). The findings reveal both positive
and negative association structures with different foreign exchange rates, gold prices,
and stock indices. The LASSO technique identifies EPU, gold returns, and search inten-
sity as the significant drivers of Bitcoin returns (Panagiotidis et al. 2018).
The literature reports various artificial intelligence and predictive modeling
approaches for forecasting future movements of financial assets. For example, a combi-
nation of radial basis function neural network (RBFN), K-means clustering, and artifi-
cial fish swarm algorithm (AFSA) estimates the future movements of the Shanghai stock
exchange (Shen et al. 2011). The model performs better than five other approaches. Sim-
ilarly, a combination of discrete wavelet transforms (DWT), MARS, and SVR predicts
stock market returns (Kao et al. 2013). This approach outperforms DWT-MARS, DWT-
SVR, ARIMA, standalone SVR, and ANFIS. Similar other hybrid models exist for stock
market prediction (Oztekin et al. 2016; Wang et al. 2016; Chen and Hao 2017). These
models also perform better compared to the traditional and some other hybrid forecast-
ing approaches. Unfortunately, they are not used to predict the future prices of Bitcoin.
Various techniques like neural network (Jang and Lee 2017), neuro-fuzzy (Atsalakis
2019), machine learning (Mallqui and Fernandes 2019), deep learning (McNally et al.
2018; Wu et al. 2018), and deep neural network (Nakano et al. 2018), deep learning
chaotic neural networks (Lahmiri and Bekiros 2019) predict the directional predictive
modeling of daily prices of Bitcoin. These techniques produce superior predictive per-
formances over their traditional counterparts and are more useful for nonlinear and cha-
otic financial markets. The literature reveals that Bitcoin is very closely associated with
a myriad of global financial assets. The multifractal interactions and spillovers of Bit-
coin with these assets make the temporal dynamics very complicated. It is challenging
to identify explanatory variables beforehand due to these associations. The degree of
(in)efficiency owing to the Bitcoin bubble and other extreme events further complicates

13
298 Annals of Operations Research (2021) 306:295–320

the predictive analysis task. Therefore, developing a robust predictive framework for
predicting future prices of Bitcoin can be of paramount importance for practical
implications.
Time-series decomposition-based predictive approaches that utilize lagged information
as explanatory features for estimating financial assets’ future movements emerge to be suc-
cessful. However, it is rare to find a granular framework-based predictive model driven by
lagged information for forecasting Bitcoin prices. The present work contributes to the exist-
ing literature by proposing a granular predictive architecture that can recognize Bitcoin
price movements’ intricate patterns. Wavelet decomposition is an instrumental paradigm
for this purpose. SVR works as the learning algorithm. Deployment of DE for optimiz-
ing SVR parameters augments the quality of the final forecasts. Wavelet decompositions
generate granular linear and nonlinear subseries. Pattern recognition algorithms apply to
these subseries for generating forecasts. The use of SVR on all the components may result
in overfitting. Thus, we apply metaheuristic driven SVR to the nonlinear component and
PRI model to the linear component to reduce this effect. The DE algorithm selects the PRI
parameters.

3 Data description

We compile daily closing prices of Bitcoin for January 10, 2013, to February 23, 2019 for
conducting the predictive exercise. During late 2017, Bitcoin prevails to be the most domi-
nant cryptocurrency with more than 42% market share capitalization. As a result, we observe
a steady upward growth of Bitcoin price during this period. Due to heavy criticism and the
Bitcoin bubble, the growth eventually declines afterward. So, we consider the study period
consisting of significantly normal, bullish, and bearish phases. Figure 1 portrays the tempo-
ral evolutionary dynamics of Bitcoin prices during the considered period. Table 1 reports the

Fig. 1  Temporal evolutionary pattern of Bitcoin

13
Annals of Operations Research (2021) 306:295–320 299

results of the statistical tests and additional properties to comprehend the underlying dataset’s
empirical properties.
Financial time series exhibits nonstationary, nonparametric, and chaotic traces (Avdou-
las et al. 2018; Guerard et al. 2020). Mere visualization of the temporal evolutionary pat-
terns may not capture such properties distinctly. Throughout the literature, researchers rely
upon statistical and econometric tests for decoding such properties (Das et al. 2018; Ghosh
et al. 2020; Jana et al. 2019; Boukhatem et al. 2020). The present work follows the same
principles for comprehending the fundamental temporal properties of Bitcoin prices.
Outcomes of Jarque–Bera and Frosini tests indicate that daily closing prices of Bitcoin
over the considered period do not abide by the normal distribution. Test statistics of Philip-
Perron and Augmented Dickey-Fuller tests emerge insignificant, indicating unit-roots’
presence in Bitcoin price dynamics. So, the time series under consideration is nonstation-
ary. Therefore, the chosen time series’s complex evolutionary pattern implies that the tra-
ditional econometric models for forecasting are not appropriate for deriving forecasts. We
perform Terasvirta’s and White’s NN tests to further evaluate the behavioral patterns for
gaining more in-depth insights. The test statistic values of both these tests are significant.
So, there exists a strong nonlinear trait in the evolutionary patterns of Bitcoin prices. The
Hurst exponent identifies the randomness and persistence of the time series. The Hurst
exponent value of 0.89404 (> 0.5) implies the presence of fractional Brownian motion with
a persistent trend (Ghosh et al. 2017). The persistent trend further indicates the presence
of volatility clustering. Thus, the immediate past information on Bitcoin price does play a
significant role in explaining future behavior. The present study exploits this property by
setting one, two, three, four, and five day back Bitcoin prices as independent variables to
predict a one-day ahead price. The partial autocorrelation plots shown in Sect. 5 validates
the selection of five lagged variables.

4 Research methodology

The proposed optimization-based regression approach has four distinct components—

MODWT, SVR, PRI, and DE. As the closing prices of Bitcoin are persistent, we select the
explanatory variables based on the lagged information. This section elucidates the details
of individual components.

4.1 MODWT

MODWT decomposes the original Bitcoin prices into granular subcomponents. It deter-
mines the respective wavelet coefficients accounting for different time horizons. Daube-
chies least asymmetric filter of length eight (la8) helps in the decomposition process.
The technique has additional advantages over the conventional ’Haar’ wavelet filters in
smoother operations (Gencay et al. 2002) and the generation of unrelated coefficients (Cor-
nish et al. 2006). MODWT segregates the Bitcoin price time series into time and frequency
space by expressing the sequence as projections of the father (𝜙) and mother (𝜓 ) wave-
lets representing approximation and units, respectively. Mean figures of approximation and
detail components are one and zero, as shown in the following equations:

∫ (1)
𝜙(t)dt = 1

13
 300

13
Table 1  Empirical properties of Bitcoin price
Jarque–Bera Test Frosini Test Philip-Perron Test Augmented Dickey-Fuller Test Terasvirta’s Test White’s NN Test Hurst Exponent

2031.86*** 5.5217*** − 9.5062# − 2.5873# 33.7355*** 23.2801*** 0.89404

#
Not Significant, ***Significant at 1% level of significance
Annals of Operations Research (2021) 306:295–320
Annals of Operations Research (2021) 306:295–320 301

∫ (2)
𝜓(t)dt = 0

We express the long-scale linear component through the father wavelet, whereas
the mother wavelet represents the linear component’s nonlinear and volatile deviations.
Mathematically, we write them as:
( )
t − 2j k
𝜙(j,k) = 2−j∕2 𝜙 (3)
2j

( )
t − 2j k
𝜓(j,k) = 2−j∕2 𝜓 (4)
2j

From the original time series, we can approximate the coefficients as follows:

∫ (5)
s(j,k) = f (t)𝜙(j,k)

∫ (6)
d(j,k) = f (t)𝜓(j,k)

The combination of wavelet functions formulates the original time series f (t) as
follows:
∑ ∑ ∑ ∑
f (t) = sJ,k 𝜙J,k (t) + dJ,k 𝜓J,k (t) + ⋯ + dj,k 𝜓j,k (t) + ⋯ + d1,k 𝜓1,k (t) (7)
k k k k

We write Eq. (7) as:

f (t) = SJ + DJ + DJ−1 + ⋯ + Dj + ⋯ + D1 (8)
where
∑
SJ = sj,k 𝜙j,k (t) (9)
k

∑
DJ = dj,k 𝜓j,k (t), j = 1,2, … , J (10)
k
[ ]
The orthogonal constituents SJ , DJ , DJ−1 , … , D1 are the outcomes of the decompo-
sition process of the original time series f (t) . DJ denotes the wavelet detail at J th level.
SJ defines the sum of deviations at the detail scales. In the proposed predictive analyt-
ics exercise, we carry out six decomposition levels that resulted in one approximation
or linear component and six detail components accounting for high frequency or non-
linearity traits of the time series. Subsequently, predictive algorithms applied to the
respective components yield predictions. For the MODWT decomposition.
The time series consisting of the closing prices of Bitcoin along with explanatory
features become inputs for the decomposition exercise. The outputs comprise of granu-
lar decomposed series of Bitcoin prices and other explanatory features.

13
302 Annals of Operations Research (2021) 306:295–320

4.2 Support vector regression (SVR)

SVR (Vapnik, 1995) is an extension of the SVM algorithm applicable for performing non-
linear regression. A nonlinear mapping function 𝜃(x) projects the input xi ∈ Rp to a feature
space with a high dimension and establishes a linear relationship between the target and
predictor constructs. Mathematically, we represent this as follows:

f (x) = wT 𝜃(x) + b (11)

where w and b are the weight and bias parameters, respectively, account for the decision
boundary.
The goal is to estimate the parameters that eventually result in at most 𝜀 deviation from
the target for all training samples. This set up of SVR is known as 𝜀-SVR.

0 if |yi − f xi | ≤ 𝜀
{ | ( )|
| ( )|
|y i − f xi | = |
( )| | (12)
| |𝜀 |
|yi − f xi | − 𝜀 Otherwise
| |
where yi denotes the target variable. We estimate f (x) by minimizing the following quad-
ratic optimization problem through determining the parameters w and b using slack con-
structs, 𝜉i , 𝜉i∗:

1 � n
Minimize ∶ ‖w‖2 + C (𝜉i + 𝜉i* ) (13)
2 i=1

subject to

⎧ wT x − y ≤ 𝜀 + 𝜉
⎨ yi − w xi ≤ 𝜀 + 𝜉 i
⎪ i i i
(14)
T ∗

⎪ 𝜉i , 𝜉i ≥ 0, i = 1, … , n
∗
⎩

where C is the regularization parameter represents the trade-off between the fitness of
regression and resultant error in training. We write the dual problem with the final solution
using Eq. (11) as follows:
∑
n
( )
f (x) = 𝛼i − 𝛼i∗ k(xi , xi∗ ) + b (15)
i=1

where k(xi , xi∗ ) is a kernel function,

� 𝛼�i and 𝛼i �are Lagrange
∗
multipliers.
� We use the Radial
basis kernel (RBF) defined by k xi , xj = exp −‖xi − xj ‖2 ∕2𝜎 2 where 𝜎 denotes the width
of the kernel. The three critical parameters C , 𝜎 , and 𝜀 that drive the learning process
and affect the prediction performance. The DE provides their optimum values. The SVR-
DE learning model applied to the nonlinear decomposed components obtained through
MODWT performs the predictive analytics task.

4.3 Polynomial regression with interaction (PRI)

PRI is an extension of simple linear regression. It fits a polynomial to establish the relationship
between the target and explanatory variables. Usually, the ordinary least square (OLS) tech-
nique estimates the regression coefficients, excluding the predictors’ interactions. These two

13
Annals of Operations Research (2021) 306:295–320 303

issues lead to the underfitting of the regression curve when time-series movements are highly
nonstationary and nonlinear, as in Bitcoin prices. We use a third-order quadratic regression
model with one level of interaction with five lagged variables as explanatory variables. The
polynomial of order three with one level of interaction is as follows:
∑
3
∑
3
j
∑
3
∑
3
∑
3
Ytest = 𝛼 + i
𝛽i Yt−1 + 𝛽j+3 Yt−2 + k
𝛽k+6 Yt−3 + l
𝛽l+9 Yt−4 + m
𝛽m+12 Yt−5
i=1 j=1 k=1 l=1 m=1

∑
4
∑
3
∑
2
∑
1
+ 𝛿n Yt−1 Yt−n−1 + 𝛿o+4 Yt−2 Yt−n−2 + 𝛿p+7 Yt−3 Yt−p−3 + 𝛿q+9 Yt−4 Yt−n−4
n=1 o=1 p=1 q=1
(16)
where Ytestrepresents the estimated Bitcoin price at time t, Yt−i accounts for the previous
information of Bitcoin price at i(i = 1,2, … , 5) period back, 𝛼 , 𝛽, and 𝛿 are the respective
coefficients of polynomial and interaction terms.
The DE estimates the regression coefficients 𝛼, 𝛽i and 𝛿i given in the Eq. (16) as follows:
�
Find ∶ 𝛼, 𝛽i , 𝛿i
∑ � act �2 (17)
Argmin ∶ Yt − Ytest

where Ytact represents actual Bitcoin price at time t.

4.4 Differential evolution

The DE is a widely used metaheuristic for solving complex optimization problems in the con-
tinuous domain (Storn and Price 1997). It is a population-based searching algorithm that trav-
erses the search space intelligently to obtain local or near-optimal solutions. The following are
the detailed procedures of individual phases of the DE.

4.4.1 Initialization
−
The initial population of the DE comprises of P vectors { xi,G,i = 1,2, … , P}, where G denotes
the generation number. We randomly generate the initial population from a uniform distri-
bution specified lower and upper bounds. The possible solutions can be represented as a
D-dimensional vector as follows:
− ( )
xi,G = xi,1,G , xi,2,G , … , xi,D,G (18)

The selection, recombination, and mutation operations subsequently modify the initial
population.

4.4.2 Mutation
(−)
In traditional DE, three vectors estimate the mutation vector v . Mathematically, we may
represent it as follows:
− − (− − )
vi,G = xr1 ,G + F ∗ xr2 ,G − xr3 ,G (19)
−
where vi,G is the donor vector, and F is the scaling parameter that varies between 0 and 2.

13
304 Annals of Operations Research (2021) 306:295–320

4.4.3 Recombination

In this phase, the components of a donor vector generate a trial offspring vector as
follows:

vi,j,G , if randj ≤ CR or j = rn(j)

{
ui,j,G =
xi,j,G , otherwise (20)

where randj (j = 1,2, … , D) denotes random numbers between 0 and 1; rn(j) denotes a ran-
domly selected index from 1,2, … , D and CR is the crossover parameter between 0 and 1.

4.4.4 Selection

The DE uses a greedy approach for selecting candidate vectors for the subsequent gen-
erations. It compares trial and target vectors and makes the final selection as follows:
{ (− ) (− )
− ui,G , if f ui,G < f xi,G
xi,G+1 = − (21)
xi,G , otherwise

The number of generations can be set depending upon the problem’s nature or fixed
until the convergence in solutions is noticed.
The DE tunes the hyper-parameters of the SVR model. It randomly initializes the
values of different parameters as input and traverses the search space intelligently to
yield optimal input parameters. In the PRI model, the DE randomly initializes the value
of the coefficients. After completion of the search process, it provides the optimal val-
ues of the same as the output.
The combination of SVR-DE and PRI-DE predictions generate the final fore-
cast of the next-day Bitcoin price. The SVR-DE approach yields granular forecasts
̂tdi (i = 1,2, … , 6) from disentangled nonlinear components provided by MODWT. The
Y
PRI-DE approach produces granular forecasts Y ̂ts6 from the approximation component
fetched by MODWT. We write the final forecast ( Y ̂ Final ) as follows:
t

̂td1 + Y
̂ Final = Y
Y ̂td3 + Y
̂td2 + Y ̂td5 + Y
̂td4 + Y ̂td6 + Y
̂ts6 (22)
t

4.5 Predictive performance indicators

Mean squared error (MSE) measures the fitness of the SVR-DE model as follows:
N ( )2
1 ∑ ̂
MSE = Yt − Yt (23)
N t=1

where, Ŷt and Yt are predicted and actual figures. The PRI-DE model considers Eq. (17) as
the fitness function.
Nash–sutcliffe efficiency (NSE) NSE measures the residual variance’s relative
strength resulting from a predictive model with the original variance of the dataset.

13
Annals of Operations Research (2021) 306:295–320 305

∑N � ̂ �2
t=1
Yt − Yt
NSE = 1 − � �2 (24)
∑N −
i=1 Yt − Yt

The NSE values lie between (−∞) to 1. The NSE value close to 1 implies excellent pre-
dictive performance by the respective model.
Index of agreement (IA) IA is a measure that considers the magnitude of error generat-
ing from a model to assess the predictive capability.
∑N 2
̂t − Yt )
t=1
(Y
IA = 1 − � �2
∑N � −� − (25)
�Ŷ �
i=1 � t − Yt � + �Yt − Yt �
� �
Like NSE, the IA figure should ideally be close to 1 for efficient forecasting.
Theil Index (TI) We can compute the index as follows:
� � �2 �1∕2
1 ∑N ̂
N t=1 Yt − Yt
TI = � �1∕2 � (26)
∑N � ̂ �2 ∑ � �2 �1∕2
1
N i=1
Yt + N1 Ni=1 Yt

Unlike NSE and IA, TI values need to be close to 0 to imply a predictive model’s
effectiveness.
Directional predictive accuracy (DA) For calculating the DA, we use the following
relationships:
{ ( )
1 ∑
N
1, (Yt+1 − Yt ) Ŷt+1 − Yt ≥ 0
DA = D ,D = (27)
N t=1 t t 0, Otherwise

The DA values near 1 indicate high accuracy of directional prediction obtained by the
model, while DA figures near 0 imply the opposite scenario. As the present paper deals
with regression, DA’s computation eventually evaluates the presented forecasting frame-
work’s directional predictive capacity. Models having high DA values are attractive for
traders for profitable trading.
Apart from self-assessment of the proposed forecasting approach’s predictive perfor-
mance, the paper also carries out a comparative study with six advanced predictive mode-
ling algorithms: MLP, ANFIS, RF, SVR, MARS LASSO to justify the usage of the predic-
tion framework. We conduct the Diebold-Mariano Statistical test for inspecting the equal
predictive ability to draw inference on comparative assessment.
Diebold-Mariano (DM) test The DM test assesses the predictive accuracy of multiple
forecasting models. We use the mean square prediction error (MSPE) as the loss function
in the DM test. The DM statistic (DMS) has the following form:
̄
D
DMS = √ (28)
VD̄ ∕M

where

13
306 Annals of Operations Research (2021) 306:295–320

∑M
̄ = 1
D d (29)
M t=1 t

M (( )2 ( )2 )
1 ∑ ̂ ̂
dt = Yt − Yttest
− Yt − Ytbench
(30)
M t=1

∑
∞
VD̄ = 𝛾0 + 2 𝛾q (31)
q=1

( )
𝛾q = cov dt , dt−q (32)
̂ test and Y
Y ̂ bench are predicted values achieved by test and benchmark models. Figure 2 dis-
t t
plays the proposed research methodology.
The decomposition-based predictive model utilizes two learning methods—SVR-DE
and PRI-DE. The DE tunes the hyper-parameters of the SVR model and estimates the coef-
ficients of the PRI model.

5 Results and discussions

We evaluate the proposed approach’s performance based on the original Bitcoin price
series and five surrogate time series randomly generated through the Monte-Carlo tech-
nique from the original Bitcoin price series. The surrogate series are more chaotic and ran-
dom. Therefore, the performance evaluation based on these series is essential to ascertain
the robustness and capability of the proposed forecasting approach. The parameter settings
remain constant for both cases.

5.1 Performance evaluation based on the original time series

The proposed framework utilizes one-day, two-day, three-day, four-day, and five-day back
lagged closing prices of Bitcoin as explanatory features. The partial autocorrelation func-
tion (PACF) plot (Fig. 3) justifies the selection.
The first local minima of the PACF plot occurs at lag 5. Subsequently, it exhibits grad-
ual upward and downward movements. The estimated Hurst exponent presented in Table 1
indicates that the series abide by the persistence trend. Therefore, we select the lag at which
the PACF value initially reaches close to 0 as the number of lagged variables to form the
independent feature set. Then, MODWT applies to the Bitcoin price dataset and explana-
tory features for six levels of decomposition. Figure 4 displays the original and decom-
posed series of Bitcoin prices.
This study utilizes the methodology presented by Bou-Hamad and Jamali (2020) to
evaluate in-sample (training) and out-of-sample (validation) forecasting performances. We
conduct static one-step ahead and dynamic multiple step-ahead forecasting schemes. For
both cases, the validation period is the last 252 observations that reflect the movements
of Bitcoin prices of the previous year of the considered sample. A rolling window of 747
observations in the static forecasting scheme estimates the one-day ahead closing price.

13
Annals of Operations Research (2021) 306:295–320 307

Fig. 2  Proposed research methodology

The considered sample has a total of 1959 observations. To predict the price of 748th
observation, samples spanning from observations 1 to 747 constitute the training period’s
window. The rolling window omits the oldest information and adds the latest observation
while traversing from training to validation segments. Training samples 3 to 749 generates
the forecast for 750th observation. Similarly, training samples 13 to 759 generates the fore-
cast for 760th observation in the one-step-ahead forecasting framework.
In a multi-step-ahead scheme, the predictive framework uses the entire in-sample
segment for training. Then, we partition the decomposed dataset randomly into training

13
308 Annals of Operations Research (2021) 306:295–320

Fig. 3  PACF plot

Fig. 4  MODWT decomposition of Bitcoin price

(80%) and test (20%) data. The SVR-DE and PRI-DE generate forecasts from the respec-
tive components. The DE algorithm estimates the SVR hyperparameters. The number
of iterations and the population size are 200 and 20, respectively. Figure 5 depicts the

13
Annals of Operations Research (2021) 306:295–320 309

Fig. 5  Iteration wise performance of the DE in SVR-DE model

performance of the DE in estimating the three parameters C , 𝜎 , and 𝜀 . We notice that

the DE algorithm converges while determining the three parameters. The DE also esti-
mates the coefficients of the PRI-DE model. As the number of parameters is signifi-
cantly higher in the PRI model, the number of iterations and population size is 400 and
80, respectively. Figure 6 displays a part of the parameter estimation process betraying
iteration wise details for a few selected parameters.
The presence of convergence is apparent. After completing parameter estimation and
component-wise predictions, we compute the final one day ahead forecast of Bitcoin
price using Eq. (20). Table 2 reports four performance indicators’ estimated values on
out-of-sample data points in static and dynamic setups.
The NSE and IA values of out-of-sample segments in static and dynamic cases are
substantially high and close to 1 and show the supremacy of the predictive model. The
TI values are low as well, which is ideal for a predictive model. Lastly, the DA statistic
is close to 1 for the training and test dataset, which signifies the precise one-day ahead
of Bitcoin price predictions. The model has noteworthy performance in estimating the
change of price direction of Bitcoin. Hence, the forecasting framework can effectively
predict the rise and fall movements, enabling investors to exploit Bitcoin for profitable
trading in both short and long terms. The performance of static forecasting is marginally
better than that of dynamic forecasting. Thus, the accuracy of the immediate future fore-
cast is comparatively more than the long run futuristic projections. Figure 7 exhibits the
actual versus predicted and fitment of actual and forecasted price on out-of-sample data.

13
310 Annals of Operations Research (2021) 306:295–320

Fig. 6  Iteration wise performance of the DE in PRI-DE model

Table 2  Forecasting performance Series IA TI DA NSE

of the proposed framework
Static forecasting
Bitcoin price 0.999703 0.027424 0.989514 0.998841
Dynamic forecasting
Bitcoin price 0.999542 0.028139 0.988408 0.998593

Both the figures further validate the claims inferred from the values of four perfor-
mance indicators. The original price and predicted price display very little discriminability
between the actual and forecasted. Table 3 summarizes the key parameters of six compet-
ing models.
We generate forecasts from all the competing models in a univariate setup by keeping
the same training and test dataset. Subsequently, we calculate the four-performance metrics
for comparison. Table 4 furnishes the comparative assessment findings of the proposed
approach with the six competing algorithms on test segments of static and dynamic fore-
casting settings.
The proposed model outperforms the rest of the models as NSE, IA, and DA figures
are higher, while the TI is lower than their counterparts. The competing models’ forecasts
are above the satisfactory level as the NSE, IA, and DA values are higher than 0.9, and
the TI value is considerably low. The performance of competing models degrades in the

13
Annals of Operations Research (2021) 306:295–320 311

Fig. 7  Actual and predicted Bitcoin price visualization

Table 3  Parameters of competing Model Key parameters

models
MLP Learning rate = 0.7, momentum = 0.9, number of
hidden nodes = 20
ANFIS Membership function = Gaussian
Rejected
RF Number of trees = 500, features for branching = 2
SVR C=3.5, 𝜎=3.2, 𝜀=0.1
MARS Degrees of interaction = 1, generalized cross
validation penalty per knot = 2
LASSO Lambda (regularization) = 0.3

multi-step ahead dynamic forecasting. The DM derives the statistical significance and
gains more in-depth insights into the competing models’ performance. Table 5 shows the
outcome of the test.
The DM test is sensitive to the order of constituents of a pair. A positively significant
test statistic value implies the statistical superiority of the second model over the first

13
312 Annals of Operations Research (2021) 306:295–320

Table 4  Forecasting performance evaluation

Series Proposed MLP ANFIS RF SVR MARS LASSO

Static forecasting
IA 0.999703 0.991793 0.992093 0.992957 0.992777 0.988731 0.990769
TI 0.027424 0.055312 0.044386 0.039149 0.038274 0.048718 0.047633
DA 0.989514 0.971166 0.973414 0.980486 0.981639 0.962394 0.967938
NSE 0.998841 0.998020 0.994762 0.995152 0.994953 0.987845 0.983132
Dynamic forecasting
IA 0.999542 0.991766 0.992069 0.992944 0.992759 0.988716 0.990734
TI 0.028139 0.055327 0.044405 0.039162 0.038283 0.048728 0.047669
DA 0.988408 0.971154 0.973398 0.980474 0.981627 0.962367 0.967917
NSE 0.998593 0.998007 0.994741 0.995127 0.994938 0.987823 0.983113

Table 5  The DM test results

Models MLP (1) ANFIS (1) RF(1) SVR (1) MARS (1) LASSO (1) Proposed
(1)

Static forecasting
MLP −
(2)
ANFIS 0.217# −
(2)
RF (2) 0.209# 0.216# −
SVR 0.211# 0.208# 0.223# −
(2)
MARS − 3.8245*** − 3.8663*** − 3.8792*** − 3.8804*** −
(2)
LASSO − 3.8587*** − 3.8935*** − 3.8956*** − 3.8967*** 0.231# −
(2)
Pro- 4.8429*** 4.8378*** 4.8249*** 4.8227*** 5.2461*** 5.2583*** −
posed
(2)
Dynamic forecasting
MLP −
(2)
ANFIS 0.198# −
(2)
RF (2) 0.237# 0.207# −
SVR 0.215# 0.224# 0.229# −
(2)
MARS − 3.8369*** − 3.8681*** − 3.8783*** − 3.8804*** −
(2)
LASSO − 3.8548*** − 3.8890*** − 3.8988*** − 3.8967*** 0.231# −
(2)
Pro- 4.8461*** 4.8451*** 4.8316*** 4.8289 *** 5.2461*** 5.2583*** −
posed
(2)

***Significant at 1% level, # Not significant

13
Annals of Operations Research (2021) 306:295–320 313

model. A negatively significant test statistic value implies the statistical superiority of the
first model over the second model. Finally, an insignificant test statistic value implies that
the performance of the models is significant different. A positive significant test statistic
associated with the proposed model against all other competing models signifies the pro-
posed model’s statistical superiority over the others. From the significance levels, we con-
clude that all other models outperform MARS and LASSO. There is no significant differ-
ence between the predictive performance of MLP, ANFIS, RF, and SVR. Thus, this test
offers the statistical validation of the efficacy and effectiveness of the proposed predictive
model. The proposed framework is statistically superior to all the six competing models for
forecasting one-step ahead and multi-step ahead Bitcoin prices.

5.2 Performance evaluation based on the surrogate time series

We evaluate the efficacy of the surrogate time series to validate and justify the proposed
framework’s deployment. The present research utilizes ’tseriesEntropy’ package of ’R’
platform for simulating the surrogate series using Sieve Bootsrap (Buhlmann, 1997). The
autoregressive (AR) model resamples the residuals with the replacement for enabling sim-
ulation. Using the ’surrogate.AR’ function, we generate five such series from the original
Bitcoin price series as depicted in Fig. 8. The figure shows that the generated five series are
comparatively more volatile than the original one.
As the cryptocurrency market is highly uncertain and chaotic due to various factors, it
becomes necessary to examine the proposed predictive model’s performance. Tables 6 and
7 show the outcome of the proposed framework’s predictive performance and competitive
models on static and dynamic forecasting environments on surrogate datasets.
The performance indicator values imply that the predictive performance is more than
satisfactory. Also, the performance is not as good as that of the original Bitcoin price pre-
diction summarized in Sect. 5.1. The validation process applies to a more random and
volatile series than the original time series. So, this is an expected outcome. The findings
suggest that the proposed predictive framework can predict the future values of cryptocur-
rencies that are comparatively less stable than Bitcoin. Table 8 reports the results of the
DM test on the static and dynamic forecasting setups.
Table 8 shows that the predictive performance is like that of the original time series.
So, the proposed predictive framework is beneficial in an increasingly volatile and random
environment. The overall findings of the predictive modeling bring out several interest-
ing results. Both one-step ahead and multi-step ahead forecasting assessments indicate the
proposed model’s efficiency for estimating Bitcoin prices. Static one-step ahead forecasting
performance is marginally better than its counterparts. It is comparatively easier to predict
the future figures in the short duration with better accuracy than in the long term. One-step
ahead forecasting estimates the closing price of Bitcoin on the next day. It will be useful
for daily trading exercise. In multi-step ahead dynamic forecasting, we estimate the Bit-
coin price for almost one year. The overall forecast accuracy would naturally be lower in
the dynamic case. However, the values of performance indicators suggest that the obtained
predictions are very accurate. This will help in long-term portfolio realignment, hedging,
etc.
Kristjanpoller et al. (2020) and Zhang et al. (2020) study the dynamic interplay of
Bitcoin with other assets. Paule-Vianez et al. (2020) report the safe heaven properties
of Bitcoin. Kristoufek et al. (2018) evaluate the inefficiency of the Bitcoin market.
However, these works barely throw any light on developing implementable practical

13
314 Annals of Operations Research (2021) 306:295–320

Fig. 8  Simulated surrogate series

predictive architectures capable of predicting short and long-run movements. Interest-

ingly, the literature manifests the multifractal nature of interplay (Kristjanpoller et al.
2020) and sensitivity to volatility contagion (Zhang et al. 2020) that triggers doubt on
the predictability of the Bitcoin price with a high degree of precision in the long-range
scale. The proposed approach shows that we can predict the Bitcoin price with utmost
precision using lagged information of past prices in static and dynamic environments
accounting for short and long duration forecasting.
Thus, despite the uncertain nature of Bitcoin prices (Paule Vianez et al. 2020), the
proposed approach can model its evolutionary patterns in an efficacious way. Kris-
toufek et al. (2018) find the inefficiency structure of Bitcoin in the short run. How-
ever, the outcomes of the proposed predictive exercises justified that the inefficiency
prevailed for long run time horizons. Predictive accuracy of surrogate series rational-
izes proposed architecture’s effectiveness in modeling Bitcoin price subject to extreme
external shocks. As a future endeavor and in connection with the existing literature,
the framework can examine Bitcoin price movements during unprecedented events like
Covid-19 pandemic with a lagged based architecture.

13
Annals of Operations Research (2021) 306:295–320 315

Table 6  Performance on Series IA TI DA NSE

surrogate series in static
forecasting Series one
Proposed 0.97985 0.035044 0.96535 0.95878
MLP 0.97082 0.061623 0.95237 0.95532
ANFIS 0.97166 0.062382 0.94961 0.95549
RF 0.97231 0.059831 0.95837 0.95731
SVR 0.97175 0.063208 0.95318 0.96108
MARS 0.96898 0.066013 0.94227 0.95116
LASSO 0.96926 0.065744 0.94359 0.95184
Series two
Proposed 0.96389 0.037921 0.93982 0.94337
MLP 0.95443 0.064318 0.92096 0.93910
ANFIS 0.95418 0.063248 0.92262 0.93886
RF 0.95562 0.062925 0.92126 0.93954
SVR 0.95380 0.062475 0.92278 0.93902
MARS 0.95085 0.065149 0.91867 0.93626
LASSO 0.95177 0.065022 0.91920 0.93581
Series three
Proposed 0.94214 0.040185 0.91887 0.91684
MLP 0.92838 0.077623 0.90472 0.91123
ANFIS 0.92769 0.078014 0.90526 0.91158
RF 0.92857 0.076352 0.90516 0.91209
SVR 0.92711 0.077484 0.90569 0.91202
MARS 0.92430 0.080169 0.90291 0.90943
LASSO 0.92461 0.080074 0.90327 0.90959
Series four
Proposed 0.95638 0.038772 0.90697 0.93058
MLP 0.94344 0.069585 0.88701 0.92155
ANFIS 0.94275 0.069192 0.89141 0.92119
RF 0.94393 0.068713 0.89176 0.92208
SVR 0.94346 0.069626 0.88637 0.92089
MARS 0.94057 0.071233 0.88205 0.91865
LASSO 0.94096 0.070876 0.88289 0.91887
Series five
Proposed 0.94196 0.039561 0.90875 0.92658
MLP 0.92768 0.072736 0.89912 0.91771
ANFIS 0.92689 0.072584 0.89873 0.91733
RF 0.92830 0.072371 0.89966 0.91808
SVR 0.92789 0.072905 0.89832 0.91756
MARS 0.92494 0.073688 0.89513 0.91412
LASSO 0.92463 0.073641 0.89545 0.91397

13
316 Annals of Operations Research (2021) 306:295–320

Table 7  Performance on Series IA TI DA NSE

surrogate series in dynamic
forecasting Series one
Proposed 0.96944 0.035759 0.96251 0.95129
MLP 0.95893 0.062069 0.94965 0.93986
ANFIS 0.95814 0.061940 0.94910 0.93967
RF 0.95917 0.060238 0.95103 0.94223
SVR 0.95785 0.062124 0.94815 0.94136
MARS 0.95109 0.063496 0.94223 0.93254
LASSO 0.95074 0.063455 0.94178 0.93297
Series two
Proposed 0.95031 0.038045 0.93724 0.94106
MLP 0.96389 0.066703 0.92761 0.92996
ANFIS 0.94214 0.066816 0.92691 0.92961
RF 0.95638 0.066234 0.92778 0.93207
SVR 0.94196 0.066589 0.92765 0.93128
MARS 0.94196 0.068117 0.92542 0.92766
LASSO 0.94196 0.068077 0.92580 0.92797
Series three
Proposed 0.92970 0.041296 0.90855 0.91522
MLP 0.91821 0.080927 0.90875 0.92658
ANFIS 0.91845 0.080918 0.90875 0.92658
RF 0.91990 0.080335 0.90875 0.92658
SVR 0.91873 0.080891 0.90875 0.92658
MARS 0.91606 0.082216 0.90875 0.92658
LASSO 0.91575 0.081995 0.90875 0.92658
Series four
Proposed 0.94373 0.038837 0.91037 0.92916
MLP 0.93236 0.070862 0.90082 0.92018
ANFIS 0.93277 0.070814 0.90055 0.92033
RF 0.93344 0.070633 0.90104 0.92105
SVR 0.93258 0.070705 0.90029 0.92046
MARS 0.92850 0.071276 0.89565 0.91914
LASSO 0.92911 0.070975 0.89533 0.91827
Series five
Proposed 0.93106 0.039840 0.91673 0.92347
MLP 0.92135 0.073084 0.90347 0.92441
ANFIS 0.92119 0.073077 0.90304 0.92429
RF 0.92202 0.072918 0.90395 0.92467
SVR 0.92168 0.073109 0.90374 0.92541
MARS 0.91825 0.073863 0.89950 0.92118
LASSO 0.91849 0.073839 0.89943 0.92093

13
Annals of Operations Research (2021) 306:295–320 317

Table 8  The DM test of predictive performance on the surrogate series

Models MLP (1) ANFIS (1) RF(1) SVR (1) MARS (1) LASSO (1) Proposed (1)

Static forecasting
MLP (2) −
ANFIS (2) 0.244# −
RF (2) 0.189# 0.226# −
SVR (2) 0.234# 0.239# 0.213# −
MARS (2) − 3.7683*** − 3.7580*** − 3.8698*** − 3.8753*** −
LASSO (2) − 3.8491*** − 3.8881*** − 3.8479*** − 3.8811*** 0.254# −
Proposed (2) 5.3628*** 5.2869*** 5.2263*** 5.3259*** 5.6780*** 5.7569*** −
Dynamic forecasting
MLP (2) −
ANFIS (2) 0.214# −
RF (2) 0.225# 0.185# −
SVR (2) 0.241# 0.228# 0.242# −
MARS (2) − 3.8316*** − 3.8579*** − 3.8856*** − 3.88923*** −
LASSO (2) − 3.8641*** − 3.8958*** − 3.8911*** − 3.9104*** 0.239# −
Proposed (2) 5.6673*** 5.6389*** 5.5427*** 5.6160*** 5.8906*** 6.1425*** −

***Significant at 1% level, # Not significant

6 Conclusions

This research proposes a novel DE-based regression framework for forecasting one
day ahead price of Bitcoin. It contributes to the forecasting literature by combining a
metaheuristic search technique with regression models in a granular setup. The utiliza-
tion of the DE for optimizing the learning algorithms’ performance assists in augmenting
the forecast quality. The proposed approach emerges as statistically superior to all other
competing models on static and dynamic forecasting situations. LASSO and MARS pro-
duce comparatively less accurate forecasts. There is no statistical evidence to discriminate
RF, MLP, ANFIS, and SVR’s predictive performance. The performance of static one-step
ahead is comparatively better than the multi-step ahead dynamic forecasting. The proposed
model yields accurate forecasts on the surrogate time series and indicates its efficiency in
recognizing intricate patterns. It is useful for estimating future figures of highly volatile
cryptocurrencies and other chaotic financial assets.
The diversification benefits of Bitcoin are well known. It has the capability of providing
investment security in challenging times. These properties of Bitcoin can help in the pre-
cise estimation of its movements in short and long-run horizons. The proposed framework
is highly efficient in accomplishing this task. As a result, it will help investors in risk miti-
gation by estimating the future prices of Bitcoin. The efficacy of the proposed framework
on the surrogate series demonstrates the predictive framework’s capability to secure note-
worthy returns in the crisis period.
This research’s scope is limited to examining the predictability of the cryptocurrency
having the largest market share. The proposed framework can assess the predictability and
efficacy of other cryptocurrencies to infer the market’s behavioral characteristics at a holis-
tic level. This work provides the opportunity to offer new forecasting frameworks by inte-
grating appropriate alternate components. We may consider a new forecasting framework
by combining the ensemble empirical mode decomposition in place of MODWT. Alternate

13
318 Annals of Operations Research (2021) 306:295–320

regression approaches like regression neural network and quantile regression neural net-
work may also work in this setting. Alternate metaheuristics search techniques like PSO,
artificial bee colony, and brainstorm optimization may replace the DE. A comparative per-
formance assessment of such predictive frameworks for forecasting highly volatile and cha-
otic time series may be an interesting research topic.

Acknowledgements The authors thank the anonymous reviewers and the Main Guest Editor for their con-
structive suggestions for improving the paper.

References
Al-Yahaee, K. H., Mensi, W., & Yoon, S. M. (2018). Efficiency, multifractality, and the long-memory prop-
erty of the Bitcoin market: A comparative analysis with stock, currency, and gold markets. Finance
Research Letters, 27, 228–234.
Atsalakis, G. S., Atsalaki, I. G., Pasiouras, F., & Zopounidis, C. (2019). Bitcoin price forecasting with
neuro-fuzzy techniques. European Journal of Operations Research, 276, 770–780.
Avdoulas, C., Bekiros, S., & Boubaker, S. (2018). Evolutionary-based return forecasting with nonlinear
STAR models: evidence from the Eurozone peripheral stock markets. Annals of Operations Research,
262, 307–333.
Bariviera, A. F. (2017). The inefficiency of Bitcoin revisited: A dynamic approach. Economics Letters, 150,
6–9.
Bou-Hamad, I., & Jamali, I. (2020). Forecasting financial time-series using data mining models: A simula-
tion study. Research in International Business and Finance,. https​://doi.org/10.1016/j.ribaf​.2019.10107​
2
Boukhatem, J., Ftiti, Z., & Sahut, J. M. (2020). Bond market and macroeconomic stability in East Asia:
A nonlinear causality analysis. Annals of Operations Research. https​://doi.org/10.1007/s1047​9-020-
03519​-6
Bouri, E., Gupta, R., Lau, C. K. M., Roubaud, D., & Wang, S. (2018). Bitcoin and global financial stress: A
copula-based approach to dependence and causality in the quantiles. The Quarterly Review of Econom-
ics and Finance, 69, 297–307.
Bouoiyour, J., Selmi, R., & Wohar, M. E. (2019). Bitcoin: competitor or complement to gold? Economics
Bulletin, 39, 186–191.
Buhlmann, P. (1997). Sieve bootstrap for time series. Bernoulli, 3, 123–148.
Chaim, P., & Laurini, M. P. (2018). Is Bitcoin a bubble? Physica A: Statistical Mechanics and its Applica-
tions, 517, 222–232.
Chen, Y., & Hao, Y. (2017). A feature weighted support vector machine and K-nearest neighbor algorithm
for stock market indices prediction. Expert Systems with Applications, 80, 340–355.
Cornish, C. R., Bretherton, C. S., & Percival, D. B. (2006). Maximal overlap wavelet statistical analysis
with application to atmospheric turbulence. Boundary-Layer Meteorology, 119, 339–374.
Cretarola, A., & Figà-Talamanca, G. (2019). Detecting bubbles in Bitcoin price dynamics via market exu-
berance. Annals of Operations Research. https​://doi.org/10.1007/s1047​9-019-03321​-z
Das, D., Bhowmik, P., & Jana, R. K. (2018). A multiscale analysis of stock return co-movements and spillo-
vers: Evidence from Pacific developed markets. Physica A: Statistical Mechanics and its Applications,
502, 379–393.
Dastgir, S., Demir, E., Downing, G., Gozgor, G., & Lau, C. K. M. (2019). The causal relationship between
Bitcoin attention and Bitcoin returns: Evidence from the Copula-based Granger causality test. Finance
Research Letters, 28, 160–164.
Demir, E., Gozgor, G., Lau, C. K. M., & Vigne, S. A. (2018). Does economic policy uncertainty predict the
Bitcoin returns? An empirical investigation. Finance Research Letters, 26, 145–149.
Filho, A. C. D. S., Maganini, N. D., & Almeida, E. F. D. (2018). Multifractal analysis of Bitcoin market.
Physica A: Statistical Mechanics and its Applications, 512, 954–867.
Gencay, R., Selcuk, F., & Whitcher, B. (2002). An introduction to wavelets and other filtering methods in
finance and economics. Academic Press.
Ghosh, I., Sanyal, M. K., & Jana, R. K. (2017). Fractal inspection and machine learning-based predic-
tive modelling framework for financial markets. Arabian Journal for Science and Engineering, 43,
4237–4287.

13
Annals of Operations Research (2021) 306:295–320 319

Ghosh, I., Sanyal, M. K. and Jana, R. K. (2017a). Analysis of causal interactions and predictive modelling
of financial markets using econometric methods, maximal overlap discrete wavelet transformation and
machine learning: A study in Asian context. In: Shankar B., Ghosh K., Mandal D., Ray S., Zhang D.,
Pal S. (eds) Pattern Recognition and Machine Intelligence. PReMI 2017. Lecture Notes in Computer
Science, vol 10597. Springer, Cham.
Ghosh, I., Jana, R. K., & Sanyal, M. K. (2019). Analysis of causal interactions and predictive modelling of
financial markets using econometric methods, maximal overlap discrete wavelet transformation and
machine learning. Applied Soft Computing, 82, 105553.
Ghosh, I,Sanyal, M. K. & Jana, R. K. (2020). Co-movement and dynamic correlation of financial and energy
markets: An integrated framework of nonlinear dynamics, wavelet analysis and DCC-GARCH. Com-
putational Economics. https​://doi.org/10.1007/s1061​4-019-09965​-0.
Guerard, J. B., Xu, G., & Markowitz, H. (2020). A further analysis of robust regression modeling and data
mining corrections testing in global stocks. Annals of Operations Research. https​://doi.org/10.1007/
s1047​9-020-03521​-y
Horra, L. P., & d. l., Fuente, G. d. l. and Perote, K. . (2019). The drivers of Bitcoin demand: A short and
long-run analysis. International Review of Financial Analysis, 62, 21–34.
Jana, R. K., Tiwari, A. K., & Hammoudeh, S. (2019). The inefficiency of Litecoin: A dynamic analysis.
Journal of Quantitative Economics, 17(2), 447–457.
Jang, H., & Lee, J. (2017). An empirical study on modeling and prediction of bitcoin prices with Bayesian
neural networks based on blockchain information. IEEE Access, 6, 5427–5437.
Kao, L. J., Chiu, C. C., Lu, C. J., & Chang, C. H. (2013). A hybrid approach by integrating wavelet-based
feature extraction with MARS and SVR for stock index forecasting. Decision Support Systems, 54,
1228–1244.
Koutmos, D. (2019). Market risk and Bitcoin returns. Annals of Operations Research, 294, 453–477.
Kristoufek, L. (2018). On Bitcoin markets (in)efficiency and its evolution. Physica A: Statistical Mechanics
and its Applications, 503, 257–262.
Kristjanpoller, W., Bouri, E., & Takaishi, T. (2020). Cryptocurrencies and equity funds: Evidence from an
asymmetric multifractal analysis. Physica A: Statistical Mechanics and its Applications, 545, 123711.
Kumar, P. D., & Ravi, V. (2017). Forecasting financial time series volatility using particle swarm optimiza-
tion trained quantile regression neural network. Applied Soft Computing, 35–52.
Lahmiri, S., & Bekiros, S. (2019). Cryptocurrency forecasting with deep learning chaotic neural networks.
Chaos, Solitons & Fractals, 118, 35–40.
Mallqui, D. C. A., & Fernandes, R. A. S. (2019). Predicting the direction, maximum, minimum and closing
prices of daily Bitcoin exchange rate using machine learning techniques. Applied Soft Computing, 75,
596–606.
Matta, M., Lunesu, I., & Marchesi, M. (2016). Is Bitcoin’s market predictable? Analysis of web search and
social media. In: Fred A., Dietz J., Aveiro D., Liu K., Filipe J. (eds) Knowledge Discovery, Knowledge
Engineering and Knowledge Management. IC3K 2015. Communications in Computer and Information
Science, vol. 631. Springer, Cham.
McNally, S., Roche, J., & Caton, S. (2018). Predicting the price of bitcoin using machine learning. In 2018
26th Euromicro International Conference on Parallel, Distributed and Network-based Processing
(PDP) (pp. 339–343). IEEE.
Nadarajah, S., & Chu, J. (2017). On the inefficiency of Bitcoin. Economics Letters, 150, 6–9.
Nakano, M., Takahashi, A., & Takahashi, S. (2018). Bitcoin technical trading with artificial neural network.
Physica A: Statistical Mechanics and its Applications, 510, 587–609.
Oztekin, A., Kizilaslan, R., Freund, S., & Iseri, A. (2016). A data analytic approach to forecasting daily
stock returns in an emerging market. European Journal of Operational Research, 253, 697–710.
Panagiotidis, T., Stengos, T., & Vravosinos, O. (2018). On the determinants of bitcoin returns: A LASSO
approach. Finance Research Letters, 27, 235–240.
Paule-Vianez, P.-R., & C. and Gomez-Martinez, R. . (2020). Economic policy uncertainty and Bitcoin. Is
Bitcoin a safe-haven asset? European Journal of Management and Business Economics. European
Journal of Management and Business Economics, 29, 347–363.
Poyser, O. (2019). Exploring the dynamics of Bitcoin’s price: a Bayesian structural time series approach.
Eurasian Economic Review, 9, 29–60.
Shen, W., Guo, X., Wu, C., & Wu, D. (2011). Forecasting stock indices using radial basis function neural
networks optimized by artificial fish swarm algorithm. Knowledge-Based Systems, 24, 378–385.
Storn, R., & Price, K. (1997). Differential evolution: A simple and efficient adaptive scheme for global opti-
mization over continuous spaces. Journal of Global Optimization, 11, 341–359.
Takaishi, T. (2018). Statistical properties and multifractality of Bitcoin. Physica A: Statistical Mechanics
and its Applications, 506, 507–519.

13
320 Annals of Operations Research (2021) 306:295–320

Tiwari, A. K., Jana, R. K., Das, D., & Roubaud, D. (2018). Informational efficiency of Bitcoin: An exten-
sion. Economics Letters, 163, 106–109.
Urquhart, A. (2016). The inefficiency of Bitcoin. Economics Letters, 148, 80–82.
Vapnik, V. (1995). The nature of statistical learning theory (2nd ed.). Springer.
Wang, J., Hou, R., Wang, C., & Shrn, L. (2016). Improved v -Support vector regression model based on
variable selection and brainstorm optimization for stock price forecasting. Applied Soft Computing, 49,
164–178.
Wu, C. H., Lu, C. C., Ma, Y. F., & Lu, R. S. (2018). A new forecasting framework for bitcoin price with
LSTM. In 2018 IEEE International Conference on Data Mining Workshops (ICDMW) (pp. 168–175),
IEEE.
Wang, G. J., Xie, C., Wen, D., & Zhao, L. (2019). When Bitcoin meets economic policy uncertainty
(EPU): Measuring risk spillover effect from EPU to Bitcoin. Finance Research Letters,. https​://doi.
org/10.1016/j.frl.2018.12.028
Zhang, Y. J., Bouri, E., Gupta, R., & Ma, S. J. (2020). Risk spillover between Bitcoin and conventional
financial markets: An expectile-based approach. The North American Journal of Economics and
Finance. https​://doi.org/10.1016/j.najef​.2020.10129​6

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

13