forecasting

Article
On Forecasting Cryptocurrency Prices: A Comparison of
Machine Learning, Deep Learning, and Ensembles
Kate Murray 1 , Andrea Rossi 2 , Diego Carraro 3 and Andrea Visentin 1,2,3, *

1 School of Computer Science & IT, University College Cork, T12 XF62 Cork, Ireland
2 Centre for Research Training in Artificial Intelligence, University College Cork, T12 XF62 Cork, Ireland
3 Insight Centre for Data Analytics, University College Cork, T12 XF62 Cork, Ireland
* Correspondence: [email protected]

Abstract: Traders and investors are interested in accurately predicting cryptocurrency prices to
increase returns and minimize risk. However, due to their uncertainty, volatility, and dynamism,
forecasting crypto prices is a challenging time series analysis task. Researchers have proposed
predictors based on statistical, machine learning (ML), and deep learning (DL) approaches, but
the literature is limited. Indeed, it is narrow because it focuses on predicting only the prices of
the few most famous cryptos. In addition, it is scattered because it compares different models on
different cryptos inconsistently, and it lacks generality because solutions are overly complex and
hard to reproduce in practice. The main goal of this paper is to provide a comparison framework
that overcomes these limitations. We use this framework to run extensive experiments where we
compare the performances of widely used statistical, ML, and DL approaches in the literature for
predicting the price of five popular cryptocurrencies, i.e., XRP, Bitcoin (BTC), Litecoin (LTC), Ethereum
(ETH), and Monero (XMR). To the best of our knowledge, we are also the first to propose using the
temporal fusion transformer (TFT) on this task. Moreover, we extend our investigation to hybrid
models and ensembles to assess whether combining single models boosts prediction accuracy. Our
evaluation shows that DL approaches are the best predictors, particularly the LSTM, and this is
consistently true across all the cryptos examined. LSTM reaches an average RMSE of 0.0222 and MAE
of 0.0173, respectively, 2.7% and 1.7% better than the second-best model. To ensure reproducibility
Citation: Murray, K.; Rossi, A.; and stimulate future research contribution, we share the dataset and the code of the experiments.
Carraro, D.; Visentin, A. On
Forecasting Cryptocurrency Prices: A
Keywords: cryptocurrency prediction; time series forecasting; deep learning; machine learning;
Comparison of Machine Learning,
ensemble modelling; temporal fusion transformer; recurrent neural networks; bitcoin
Deep Learning, and Ensembles.
Forecasting 2023, 5, 196–209. https://
doi.org/10.3390/forecast5010010

Academic Editor: Vasilios 1. Introduction

Plakandaras
Cryptocurrencies are virtual currencies that rely on blockchain technology. They
Received: 5 December 2022 have seen widespread market adoption since the introduction of Bitcoin in 2009, the
Revised: 13 January 2023 most popular crypto so far. Many different subjects trade cryptos and invest in crypto
Accepted: 18 January 2023 funds and companies; according to CoinMarketCap [1], the global market capitalisation
Published: 29 January 2023 of cryptocurrencies reached an estimated value of USD 932.49 billion in September 2022.
Although investments have seen lucrative returns, ubiquitous price fluctuations across
most cryptocurrencies make such investments challenging and risky. For example, Bitcoin’s
price has been highly volatile since its market launch, reaching peaks as high as +122% and
Copyright: © 2023 by the authors.
+1360% in 2016 and 2017, respectively [2]. Ethereum, XRP, and Litecoin have seen similar
Licensee MDPI, Basel, Switzerland.
fluctuations in 2017 alone [2].
This article is an open access article
For these reasons, investors require a forecasting approach to effectively capture
distributed under the terms and
crypto price fluctuations to minimise the risk and increase their profit. Moreover, it is
conditions of the Creative Commons
possible to use volatility forecasts to estimate swings in their price, which is useful for
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
developing and analysing quantitative financial trading strategies [3]. However, similar
4.0/).
to stock price forecasting, whose market is dynamic and complex as well [4], crypto price

Forecasting 2023, 5, 196–209. https://doi.org/10.3390/forecast5010010 https://www.mdpi.com/journal/forecasting

Forecasting 2023, 5 197

forecasting is regarded as one of the most challenging prediction tasks in the financial
domain at present [5]. Most successful researchers cast this problem as an example of time
series forecasting [6–11], since the idea is to leverage historical and current price data to
predict future prices over a period of time or a specific point in the future. Time series
analysis has also been applied in weather forecasting and demand forecasting for retail and
procurement, for example.
In the literature, the application of statistical techniques is the traditional approach
for time series forecasting. Such techniques adopt statistical formulas and theories to
model and capture patterns in the time series. The most frequently employed statistical
models are the autoregressive integrated moving average (ARIMA) model and its variants,
exponential mmoothing, multivariate linear regression, multivariate vector autoregressive
model, and extended vector autoregressive model [12]. In addition, in forecasting the
future prices of cryptos, the most popular example is the ARIMA [13]. Researchers have
commonly employed this model to forecast Bitcoin prices [6,14,15]. Other models have also
been applied, such as generalized autoregressive conditional heteroscedasticity (GARCH)
models in volatility forecasting of cryptos [16,17] and diffusion processes in probabilistic
forecasting of cryptos [18].
Another research branch employs machine learning (ML) models such as stochastic
gradient boosting machines [19], linear regression, random forest, support vector ma-
chines, and k-nearest neighbours [20]. By leveraging historical data, these techniques focus
on identifying the most influential features that determine future crypto prices to boost
prediction accuracy.
A third body of work employs deep learning (DL) models to tackle crypto price
forecasting, following their recent widespread success in quantitative finance [21]. Neural
networks, recurrent neural networks (RNN) such as gated recurrent unit (GRU) and
long short-term memory (LSTM), yemporal convolutional networks (TCN), and hybrid
architectures have been applied to predict prices of Bitcoin, Ethereum, and Litecoin, for
example [7,9,11]. DL approaches are considered effective at time series forecasting because
they are robust to noise, they can provide native support for data sequences, and they can
learn non-linear temporal dependencies on such sequences [22].
Although the literature has proposed statistical, ML, and DL techniques, there is no
clear evidence of which of these approaches is superior. Indeed, the research is scattered
and lacks generality because it focuses on predicting the price of a single crypto among
a small number of the most popular cryptocurrencies (mainly Bitcoin). Moreover, the
over-complexity of the model architecture makes their adoption in a real-world scenario
very challenging because implementation, training, and predictions are expensive. Lastly,
with different datasets, pre-processing strategies, and experimental methodologies, the
approaches’ comparisons are inconsistent, the experiments are hard to reproduce, and their
findings are therefore unreliable.
The main goal of this paper is to overcome these limitations and shed light on the
effectiveness of the most popular approaches proposed in the literature so far on the
crypto price prediction task. Therefore, as a major contribution, we design a framework
for comparing widely used statistical, ML, and DL approaches in predicting the price of
five popular cryptocurrencies, i.e., Ripple (XRP), Bitcoin (BTC), Litecoin (LTC), Ethereum
(ETH), and Monero (XMR). DL networks selected include different architectures such as
convolutional neural networks, recurrent neural networks, and transformers. To the best of
our knowledge, we are also the first to propose using temporal fusion transformer (TFT)
as a DL approach to tackle crypto price prediction. In addition, we investigate the use of
hybrid models and ensembles to determine whether a combination of multiple models can
improve the accuracy of the predictions.
To overcome cryptocurrency prices’ high fluctuation and volatility, we transform
non-stationary time series into stationary ones by applying detrending. Predictive mod-
els are trained and tested on a 5-year time-window dataset we collected from online
cryptocurrency trading platforms. Our evaluation methodology spans over one year of
Forecasting 2023, 5 198

data and is incremental with monthly time windows. Results show that DL approaches
are better than ML and statistical approaches, and, for DL models, complex architec-
tures outperform less complex ones. To ensure reproducibility and stimulate future re-
search contribution, we open source the dataset and the code of the experiments (https:
//github.com/katemurraay/tsa_crt, accessed 15 January 2023), as we believe our work to
be an essential starting point for practitioners to investigate crypto price prediction.
The remainder of this paper is structured as follows: Section 2 presents the models
comparison, the data collection and preprocessing, and finally describes the experimental
methodology; Sections 3 and 4 outline the results of the experiments and discuss their
findings, respectively; finally, Section 5 draws conclusions and illustrates future plans.

2. Materials and Methods

In our framework, we assume the availability of a dataset of size m with daily interval
granularity, i.e., each dataset’s instance refers to a timestamp day ti , i ∈ (1, m), where t1 and
tm denote the earliest and the latest data points available in the dataset, respectively. We
denote with yti the value of the target variable at timestamp ti , i.e., the cryptocurrency price
to predict. We also denote with xti the features available at time ti ; xti = [yti−l , . . . , yti−1 ],
where l is the length of the window considered as input by the models. Our goal is to build
predictive models that learn a function f ( xti ) = yti , see Section 2.1 for the list of models
we employ in this study. This learning task is a typical example of univariate time series
analysis because only one variable (i.e., the crypto price, y) varies over time.
In the remainder of this section, we describe the predictive models, the data acquisition
and its preprocessing, and the experimental methodology we use to compare the models.

2.1. Predictive Models

Below we give details of the statistical, ML, DL, hybrid, and ensemble models we compare.
• Auto Regressive Integrated Moving Average (ARIMA). This is a generalisation of
the simpler ARMA model (auto regressive moving average). The traditional three-
step process of constructing ARIMA models by [13], includes model identification,
parameter estimation, and finally, the diagnosis of the simulation and its verification.
Essentially, a prediction for a yttarget value is the linear combination of the yti values
up to the ttarget timestamp and the prediction errors made for the same y xti values.
Examples of ARIMA usage include forecasting for air transport demand [23,24], long-
term earning prediction [25], and next-day electricity price prediction [26]. ARIMA
has effectively predicted BTC prices in [6,14,27].
• k-Nearest Neighbor (kNN). Originally suited for classification tasks, kNN is a non-
parametric model that has been successfully extended and employed for regression
tasks in time series analysis. To predict yttarget , the kNN calculates the k most-similar
xti values to xttarget . Then, prediction of ytarget is the weighted average of the k yti
values. The kNN model has been used in financial forecasting [28], electric market
price prediction [29], and in the prediction of Bitcoin [30].
• Support Vector Regression (SVR). Built on support vector machines for classification,
SVR enables both linear and non-linear regression. Similarly to kNN, SVR is a non-
parametric methodology introduced by [31]. SVR aims to maximise generalisation
performance when designing regression functions [32]. SVR was applied to a variety
of time series tasks such as forecasting warranty claims [32], predicting blood glucose
levels [33], and for stock predictions in the financial market [34]. Examples of SVR
usage in forecasting crypto prices can be found in [20,21].
• Random Forest (RF) Regressor. This is essentially an ensemble of decision trees,
each of which is built on a random subset of the training set. RF’s predictions are
performed by averaging the predictions of individual trees. The key benefits of RF
are its generalisation capability, and minimal sensitivity to hyperparameters [35]. RF
has been used in time series tasks for forecasting cyber security incidents [36], for the
prediction of methane outbreaks in coal mines usage [37], and for projecting monthly
Forecasting 2023, 5 199

temperature variations [35]. In the prediction of cryptos, RF has been used for BTC
forecasting in [20] and BTC, ETH, and XRP in [19].
• Long Short Term Memory (LSTM). This is a type of RNN capable of learning long-
term dependencies and, therefore, is suitable for time series analysis [38]. Although
LSTMs follow a chain-like structure similar to ordinary RNNs, in an LSTM’s repeating
module, four neural layers interact, i.e., two in the input gate, one in the forget gate,
and one in the output gate. The input gate adds or updates new information, and
the forget gate removes irrelevant information. The output gate ultimately passes
updated information to the following LSTM cell. Examples of LSTM usage can be
found in short-term travel speed prediction [39], predicting healthcare trajectories
from medical records [40], and forecasting aquifer levels [41]. The model has also been
successful for crypto price prediction [7–9].
• Gated Recurrent Unit (GRU). Although the GRU model is similar to LSTM, the former
improves upon the computational efficiency of the latter because it has fewer external
gating signals in the interpolation. Consequently, the related parameters are reduced.
GRU has been used in the short-term prediction for a bike-sharing service [42], network
traffic predictions [43], and forecasting airborne particle pollution [44]. GRU was
found in [10] to forecast the prices of BTC, ETH, and LTC successfully.
• LSTM-GRU (HYBRID). This method was proposed by Patel et al. [11] to avail of the
advantages of both LSTM and GRU. Their study indicated that this hybrid approach
effectively predicted Litecoin and Monero daily prices, for this reason we include it
herein. Combinations of LSTM and GRU have been successfully applied to predict
water prices [45].
• Temporal Convolution Network (TCN). Presented by Bai, Kolter, and Koltun [46],
TCN is a variant of the convolutional neural network architecture, and uses dilated,
causal, one-dimensional convolutional layers. TCN’s causal convolutions prevent
future data from leaking into the input. TCNs have been widely adopted in time
series forecasting. For example, TCNs can produce a short-term prediction of wind
power [47], predict just-in-time design smells [48], and forecast in stock volatility [49].
In addition, TCN was effective at forecasting weekly Ethereum prices [50].
• Temporal Fusion Transformer (TFT). Introduced by [51], the architecture of TFT is built
on the vanilla transformer architecture. TFT is one of the most recent deep learning
approaches for time series forecasting. Its design incorporates novel components such
as gating mechanisms, variable section networks, static covariates, prediction intervals,
and temporal processing. TFT has been applied in other time series tasks such as the
prediction of pH levels in bodies of water [52], flight demand forecasting [53], and
projecting future precipitation levels [54]. To the best of our knowledge, we are the
first to employ it for the crypto price prediction.
We employ the voting regressor for the ensemble, a combination of different base
inducers using the models described above. We build a total of 502 ensembles, one for each
possible combination. An ensemble’s prediction is given by averaging the predictions from
the individual models that compose the ensemble. Note that each individual model was
trained separately and independently.
In our comparison, other approaches for time series forecasting could have been
investigated, for example, functional data analysis for predicting electricity prices [55,56],
group method of data handling and adaptive neuro-fuzzy inference system for predicting
faults [57], and multi-modality graph neural network for financial time series prediction [58].
However, we limited our choice to the most popular and representative models proposed
in each category (i.e., statistical, ML, and DL) in the literature because a complete and
exhaustive comparison of time series methods is beyond the scope of this paper.

2.2. Data Collection

The data were gathered from Binance.com (https://www.binance.com/en, accessed 13
July 2022) and Investing.com (https://www.investing.com, accessed 13 July 2022) websites.
Forecasting 2023, 5 200

Binance.com is the world’s largest and most popular cryptocurrency exchange portal for
daily trading. It provides an array of features specific to cryptocurrency products which
include market information for thousands of cryptocurrencies. Investing.com acts as a
global portal for stock market information and analysis on many worldwide financial
markets. For our investigation, we selected five popular cryptocurrencies in the literature,
i.e., XRP, Bitcoin (BTC), Litecoin (LTC), Ethereum (ETH), and Monero (XMR).
The data collection process made use of the Binance API as a primary resource and
it was complemented by information retrieved from Investing.com when missing values
occurred (e.g., when the closing price of XMR was not available for a specific day). The
time frame of the collected data ranges from 1 June 2017 to 31 May 2022, i.e., five years. A
summary of the resulting datasets are reported in Table 1, and the covariates available for
the i-th instance of each dataset are the following:
• ti —the timestamp of the day;
• OPti —the opening price of the cryptocurrency at ti ;
• HPti —the highest price of the cryptocurrency at ti ;
• LPti —the lowest price of the cryptocurrency at ti ;
• yti —the target variable, i.e., the closing price of the cryptocurrency at ti (which corre-
sponds to the opening price of the following day, i.e., OPti+1 = yti ).
In this paper, we address the crypto price prediction task as a univariate time series
analysis problem, and therefore we ignore the covariates OP, HP, and LP, but they are
included in the available preprocessed dataset. We plan to consider such covariates in
future work.

Table 1. Details of the cryptos analysed in this work. All prices are in US Dollars (USD).

Market 24 h
Name Release Year Cap 1 Volume 1 Min Price 2 Max Price 2 Mean Price 2 Price SD 2

Bitcoin (BTC) 2009 393.41 45.67 1914.10 67,525.83 18,621.99 17,623.38

Etherium 2015 192.46 19.31 83.76 4807.98 1021.77 1220.11
(ETH)
Litecoin 2011 3.93 0.53 23.08 387.80 101.41 64.33
(LTC)
Monero 2014 2.71 0.09 29.20 484.00 142.39 90.43
(XMR)
XRP (XRP) 2012 22.96 0.96 0.14 2.78 0.51 0.36
1In billions of US dollars (USD). Values recorded on the 31 October 2022 from CoinMarketCap [1]. 2 In US dollars
(USD). Values relative to the collected data period, 1 June 2017 to 31 May 2022.

2.3. Data Pre-Processing

When forecasting with time series, their stationarity property is crucial for effective
modeling [5]. A time series with mean and variance that do not change over time is referred
to as stationary. On the contrary, a time series whose mean, frequency, and variance
fluctuate over time and frequently display high volatility, trend, and heteroskedasticity
is referred to as non-stationary [5]. Typically, traditional statistical forecasting methods
such as ARIMA require time series to be stationary in order to successfully capture their
properties [59]; similarly, stationarity favours learning in non-statistical models such as
the ML and DL employed in this paper [60]. For these reasons, we run the augmented
Dickey–Fuller (ADF) statistical test [61] to identify whether our datasets are stationary. The
results show that all datasets are non-stationary except the XRP dataset.
We transform our datasets into stationary datasets by applying detrending, i.e., the
process of removing the trend from a time series. In particular, we apply the differencing
transformation, the simplest detrending technique that generates a new time series where
the new value y0ti at timestamp ti is calculated as the difference between the original
observation and the observation yti−1 at the previous time step, i.e.,

y0ti = yti − yti−1 (1)

Forecasting 2023, 5 201

Figure 1 shows the original Bitcoin time series in yellow and its differenced version in
red. The ADF test computed on the detrended datasets confirm their stationarity.
Another typical pre-processing step that is widely adopted to enhance learning is data
normalisation (e.g., [11]). We apply the Min-Max normalisation to all yti of each dataset, so
that values are mapped in the (0, 1) range according to the following formula:

yti − ymin
y ti = (2)
ymax − ymin

where ymin = min{yti } and ymax = max {yti }. To avoid leakage, ymin and ymax values are
calculated from training data only.

Figure 1. Bitcoin’s daily closing price from June 2017 to May 2022. We plot the original time series in
yellow and the detrended one in red.

2.4. Experimental Methodology

We performed experiments on each dataset/crypto separately, with the following
methodology that was the same for all models. We performed an initial temporal training-
test split on each dataset. The first 80% of the data belonged to the training set (i.e., four
years of data, from t = 1 June 2017 to t = 31 May 2021) and the last 20% of the data
belonged to the test set (i.e., one year of data, from t = 1 June 2021 to t = 31 May 2022). We
further partitioned the test set into twelve non-overlapping monthly windows (from June
2021 to May 2022 included) and we labelled them with Mi , i ∈ {1, 2, . . . , 12}.
Inspired by [62], an incremental monthly-based strategy was employed to evaluate
each model. In the first evaluation step, we trained the model on the training set, we
performed predictions, and we computed the test metrics (presented in Section 2.5) on M1 .
In the second evaluation step, we included M1 ’s data in the training set and we retrained
the model from scratch on this newly enlarged training set. We again performed predictions
and we computed the test metrics on M2 . We repeated the same process for the remaining
ten partitions, each time increasing the training set and moving the evaluation window one
step forward. Both ML and DL models have hyperparameters; therefore, we tuned them
only in the first evaluation step by using 20% of the training data for validation (optimizing
for MSE), and we kept them fixed for the remainder of the evaluation. Hyperparameter
details and values spaces are reported in Appendix A Table A1. We considered a sliding
window of 30 days of data as input to compute a one-step-ahead prediction. To avoid
Forecasting 2023, 5 202

overfitting of the DL models during training, we applied early stopping and we performed
the experiments three times (averaging the results) to account for the randomness in the
initialisation of the models.

2.5. Evaluation Metrics

To assess the quality of a model’s predictions, we computed the root mean squared
error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and
R-squared score (R2 ) in each evaluation step described in the previous section, as follows:
s
∑in=1 (yti − ŷti )2
RMSE = (3)
n

n
1
MAE =
n ∑ |yti − ŷti | (4)
i =1
n |yti − ŷti |
1
MAPE =
n ∑ | t ti |
∗ 100 (5)
i =1

∑in=1 (yti − ŷti )2

R2 = 1 − (6)
∑in=1 (yti − y)2
In the above Equations (3)–(6), yti is the true price of the crypto after the normalisation,
ŷti is the predicted value, y is the average of the predicted values, and n indicates their
number. Note that the R-squared metric highlights the model’s variance in relation to the
total variance. Therefore, as opposed to the other error metrics, the higher the R-squared
value, the better the model’s performance.

3. Results
This section reports the results of the experiments and compares the regression models
in terms of accuracy and computational time. We assess both their average and crypto-
specific performances. Then, we examine the results of the ensembles and the contribution
of each individual model to an ensemble’s performance.

3.1. Individual Models

Table 2 shows the average performance of each model computed across all cryptos.
Models are ranked by RMSE in ascending order.
First, we observe that the models’ ranking is consistent across all the accuracy metrics
(with very few exceptions). The LSTM exhibits the best performance, with a consistent gap
compared to the other models. For each metric, values are quite close because we compute
them on the normalised predicted price, and not on the detrended data. The recurrent
neural network models occupy the first three positions of the rank, followed by the KNN
and the convolutional network approach. Interestingly, ARIMA performs better than TFT,
RF, and SVR.
Regarding the time required to train and deploy the models, DL approaches are more
expensive compared to machine learning and statistical methods, as expected. Overall,
all the models provide a prediction in a reasonably short time, so they might be suited to
operate in some online settings. In particular, for training and inference, HYBRID (LSTM-
GRU Hybrid in Section 2.1) and TFT are the most expensive, respectively. In contrast,
ML models are considerably faster to run. The KNN provides a good trade-off between
accuracy and computational cost.
Forecasting 2023, 5 203

Table 2. The average performance of individual models ranked by RMSE in ascending order.

Model RMSE MAE MAPE R2 Train (s) Inference

(ms)
LSTM 0.02224 0.0173 3.862% 0.735 173.765 1.862
GRU 0.02285 0.0176 3.939% 0.720 254.520 1.550
HYBRID 0.02295 0.0177 3.959% 0.717 461.967 2.383
KNN 0.02332 0.0179 4.003% 0.711 <0.01 0.074
TCN 0.02334 0.0180 4.021% 0.711 40.475 1.219
ARIMA 0.02343 0.0180 4.010% 0.708 4.035 0.109
TFT 0.02353 0.0181 4.062% 0.707 105.913 8.842
RF 0.02402 0.0184 4.095% 0.697 2.121 0.586
SVR 0.02452 0.0189 4.240% 0.681 <0.01 0.008

Table 3 indicates the RMSE results across the different cryptos. The ranking of the top
three models is consistent across all the cryptos. However, in the lower positions, some
variability can be observed, e.g., SVR and TFT perform particularly well on BTC.

Table 3. The RMSE performance of individual models for each crypto (ranks are reported in brackets).

BTC ETH LTC XMR XRP Average

LSTM 0.0239 (1) 0.030 (1) 0.0189 (1) 0.0236 (1) 0.0148 (1) 0.0222 (1)
GRU 0.0245 (2) 0.0309 (2) 0.0193 (2) 0.0243 (2) 0.0153 (2) 0.0229 (2)
HYBRID 0.0246 (3) 0.0309 (3) 0.0195 (3) 0.0244 (3) 0.0154 (3) 0.0230 (3)
KNN 0.0249 (6) 0.0319 (5) 0.0197(4) 0.0245 (5) 0.0155 (4) 0.0233 (4)
TCN 0.0250 (7) 0.0319 (4) 0.0198 (5) 0.0245 (6) 0.0156 (5) 0.0233 (5)
ARIMA 0.0251 (8) 0.0320 (7) 0.0198 (6) 0.0244 (4) 0.0158 (7) 0.0234 (6)
TFT 0.0249 (5) 0.0319 (6) 0.0199 (7) 0.0250 (7) 0.0159 (8) 0.0235 (7)
RF 0.0266 (9) 0.0332 (8) 0.0199 (8) 0.0251 (8) 0.0157 (6) 0.0240 (8)
SVR 0.0248 (4) 0.0342 (9) 0.0207 (9) 0.0268 (9) 0.0160 (9) 0.0245 (9)

3.2. Ensembles
Table 4 highlights the performances of the best ten ensembles in terms of RMSE.
The ensembles do not outperform the LSTM network, and the latter is included in all
the top-performing ensembles. It is interesting to see how the LSTM and GRU ensemble
outperforms the HYBRID model, which is a deep non-sequential network that combines
LSTM and GRU.

Table 4. Average ensemble performance against individual models ranked by RMSE in ascending
order.

Ensemble RMSE MAE MAPE R2

LSTM 0.0222 0.0173 3.86% 0.73
GRU, LSTM 0.0225 0.0174 3.89% 0.73
HYBRID, LSTM 0.0225 0.0174 3.89% 0.73
HYBRID, GRU, LSTM 0.0226 0.0175 3.90% 0.73
LSTM, KNN 0.0227 0.0175 3.92% 0.73
GRU, LSTM, KNN 0.0227 0.0176 3.91% 0.72
GRU, LSTM, TCN 0.0227 0.0176 3.92% 0.72
LSTM, TCN 0.0227 0.0176 3.93% 0.72
HYBRID, LSTM, KNN 0.0227 0.0175 3.92% 0.72
HYBRID, GRU, LSTM, KNN 0.0227 0.0175 3.91% 0.72

To evaluate the contribution of an individual model, we compared the average accu-

racy of all the ensembles that include this model and those that do not (and the difference
can be seen as the average RMSE contribution given by that individual model). The results
in Table 5 confirm the individual model ranking in Table 2. Most notably, the contribu-
tions of the non recurrent models are negative, i.e., they worsen the ensemble accuracy
on average.
Forecasting 2023, 5 204

Table 5. Each model’s contribution within the ensemble ranked by difference in descending order.

RMSE without
Model RMSE with Model Model Difference (%) 1
LSTM 0.023 0.0233 1.26%
GRU 0.0231 0.0232 0.57%
HYBRID 0.0231 0.0232 0.48%
KNN 0.0232 0.0232 −0.03%
TCN 0.0232 0.0232 −0.06%
ARIMA 0.0232 0.0231 −0.2%
TFT 0.0232 0.0231 −0.21%
RF 0.0232 0.0231 −0.41%
SVR 0.0233 0.0231 −0.87%
1 (RMSE Without Model−RMSE with Model)
Difference = RMSE Without Model × 100.

4. Discussion
The results show that the models’ performance ranking is consistent across different
cryptos, and their average performance confirms the ranking. Recurrent DL approaches
dominate the cryptocurrency price prediction task according to all accuracy metrics. In
particular, the LSTM is the best-performing model with an average RMSE of 0.0222 and
substantially outperforms other network architectures, such as TCN (convolutional) and
TFT (transformer), which have a 4.9% and 5.8% higher error, respectively. The nature of
the latter architectures can explain their poor performance. Regarding TCN, convolutional
networks are good at interpreting repeated hierarchical patterns in the data (captured by
the dilated convolutions), but these patterns are absent from the crypto price time series.
Moreover, TCN generally performs better for fine-grained (dense) predictions (such as
hourly predictions rather than daily or monthly predictions). This is because the oscillation
between a wider time window has a different distribution and is harder to capture by
dilated convolutions. Regarding TFT, its attention mechanism is known for capturing
the relationship between covariates of the time series at hand. However, such covariates
are ignored in our experiments (and we leave this for future work). TCN and TFT are
also known to be data-hungry, i.e., they require substantial volumes of data to capture
patterns successfully. Unfortunately, the amount of historical data available to train these
models on forecasting daily prices is limited. The second best model is GRU, a recursive
network simpler than LSTM, which achieves an RMSE of just 2.7% higher with a similar
computational effort. To wrap up, results for DL models suggest that more expensive and
complex architectures may be redundant for this type of time series task.
The KNN provides an excellent trade-off between the accuracy of the prediction and
the computational effort required, with an error 4.8% higher than LSTM but with no training
time required and a 25 times faster inference time. The other machine learning models
(SVR and RF) are at the bottom of the ranking and, quite surprisingly, are outperformed by
the baseline ARIMA. This is probably because they cannot capture meaningful patterns in
the time series, which is noisy and presents outliers (SVR performs better because it is less
prone to outliers). In contrast, due to its linearity assumptions, ARIMA’s predictions are
directional and more accurate for short-term analysis. In conclusion, ARIMA provides a
good trade-off between good accuracy and reduced computational demand.
Ultimately, the last part of the experiment highlights that combining different regres-
sors into an ensemble does not boost performance. This approach aims to compensate for a
model’s shortcomings by averaging it with others that are more accurate in particular cases.
However, if a regressor provides more accurate predictions in the vast majority of cases,
averaging it with considerably more inaccurate models negatively affects its performance.
Indeed, the LSTM consistently outperforms all the ensembles due to a wide accuracy gap
with the other models.

5. Conclusions
This paper compares deep learning (DL), machine learning (ML), and statistical mod-
els for forecasting the daily prices of cryptocurrencies. Our one-step-ahead evaluation
Forecasting 2023, 5 205

framework is incremental and works on a monthly retraining schedule. We tested over

12 months of data. Results show that, in general, recurrent DL approaches are the best
models for this task. In particular, the LSTM is the best-performing model, and its training
is less expensive than the other DL models with the closest performance. The reasons
why DL models such as TCN and TFT underperform might be, for example, that the
convolutional approaches are better suited for dense predictions (“sparse” in our analysis)
and TFT are good at leveraging covariates (ignored in our analysis), while both approaches
suffer from a data scarcity problem. KNN and ARIMA provide a good trade-off between
accuracy and computational expense. Finally, the deployment of ensemble approaches is
detrimental, as their performance is inferior to the individual LSTM approach.
The availability of accurate predictions is essential to crypto traders, who often trade
hourly and daily. Therefore, tailoring accurate predictors for trading strategies might help
them increase their revenue. However, our predictors can only predict daily prices; in
the future, we aim to build predictors that also provide hourly prices and investigate the
integration of such predictors with some trading strategies (e.g., [3]).
Several factors can also affect the price fluctuations of cryptos, including regulations,
social media trends, market sentiments, and other cryptos’ volatility. For example, the work
of [63] analyses how regulatory news and events affect returns in the cryptocurrency market
using an event-based approach. According to this report, events that raise the likelihood
of regulation adoption are linked to a negative return for cryptos. Another example is
from [64], where the prices of other cryptos exhibit an interdependent relationship (Bitcoin
is the parent coin for both Litecoin and Zcash). Therefore, in the future, we aim to integrate
these kind of covariates in our models to improve prediction accuracy.
Another avenue of improving forecasting involves investigating the relationship
between cryptos. Their prices exhibit an interdependent relationship, and the coins can
be grouped into clusters of similar behaviour [65]. Using this framework, similar cryptos
can be used to train a more accurate model specific to that pattern and offer rich and
valuable insights into the dynamics between cryptos, while also improving the accuracy of
predictions of crypto forecasting.

Author Contributions: Conceptualisation, A.V. and A.R.; methodology, A.V., A.R., D.C. and K.M.;
software, A.R. and K.M.; validation, A.V., A.R. and K.M.; formal analysis, A.R. and K.M.; re-
sources, A.V.; data curation, A.R. and K.M.; writing—original draft preparation, D.C. and K.M.;
writing—review and editing, A.V., A.R., D.C. and K.M.; visualisation, K.M.; supervision, A.V., A.R.
and D.C.; project administration, A.V. and D.C.; funding acquisition, A.V. All authors have read and
agreed to the published version of the manuscript.
Funding: This publication has emanated from research supported in part by Science Foundation
Ireland under grant no. 18/CRT/6223. This publication has also emanated from research conducted
with the financial support of Science Foundation Ireland under grant number 12/RC/2289-P2 which
is co-funded by the European Regional Development Fund. For the purpose of Open Access, the
author has applied a CC BY public copyright licence to any Author Accepted Manuscript version
arising from this submission.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data used in the experimentation are accessible from the following
link: https://github.com/katemurraay/tsa_crt/tree/kmm4_branch/saved_data, accessed on 13
July 2022. These data were originally sourced from https://www.binance.com/ and https://www.
investing.com/, both accessed on 13 July 2022.
Conflicts of Interest: The authors declare no conflicts of interest.

Appendix A. The Hyperparameter Values of the Predictive Models

The details regarding the values of hyperparameters of each model are shown in
Table A1.
Forecasting 2023, 5 206

Table A1. Hyperparameters and architecture of forecasting models.

Model Python Library Architecture Hyperparameters Used

• convolutional layer: 64
• convolutional activation: ReLU
Single
• convolutional kernel: 5
LSTM TensorFlow Convolutional
• lstm layer : 75
Layer and a
• dense layer : 16
LSTM Layer.
• dense layer activation: ReLU
• learning rate: 1 × 10−4

• gru layer size: 75

Single GRU • gru layer activation: ReLU
GRU TensorFlow Layer and • dense layer size: 100
Dense Layer. • dense layer activation: ReLU
• learning rate: 1 × 10−3

• first lstm layer : 75

• first dropout rate: 0.05
• second lstm layer : 50
Two LSTM • lstm activation: ReLU
LSTM-GRU TensorFlow Layers and a • first dense layer : 32
Hybrid
GRU Layer. • gru layer : 50
• second dropout rate: 0.0
• second dense layer : 64
• learning rate: 1 × 10−3

• convolutinal filters: 32
• convolutional kernel: 16
Four
• dilation rate: 8
TCN TensorFlow Convolutional
• dense layer dimensions: 64
Layers
• dropout rate: 0.05
• learning rate: 1 × 10−4

• input chunk length: 30

• output chunk length: 1
• number of LSTM layers: 3
TFT DARTS • number of attention heads: 7
• hidden layer dimensions: 64
• dropout rate: 0.05
• learning rate: 1 × 10−3

• number of estimators: 200

• criterion: mse
RF Scikit-Learn • max depth: 100
• max features: sqrt
• bootstrap: True

• kernel: poly
• degree: 5
SVR Scikit-Learn • gamma: auto
• tol: 0.001
• C: 100

• number of neighbours: 28
kNN Scikit-Learn • weights: uniform
• algorithm: brute
• p: 2

• p: 1
ARIMA StatsModel • d: 0
• q: 2
Forecasting 2023, 5 207

References
1. Cryptocurrency Prices, Charts and Market Capitalizations. Available online: https://coinmarketcap.com/ (accessed on 25
November 2022).
2. Bouri, E.; Shahzad, S.J.H.; Roubaud, D. Co-explosivity in the cryptocurrency market. Financ. Res. Lett. 2019, 29, 178–183.
[CrossRef]
3. Fang, F.; Ventre, C.; Basios, M.; Kanthan, L.; Martinez-Rego, D.; Wu, F.; Li, L. Cryptocurrency trading: A comprehensive survey.
Financ. Innov. 2022, 8, 13. [CrossRef]
4. Ariyo, A.A.; Adewumi, A.O.; Ayo, C.K. Stock Price Prediction Using the ARIMA Model. In Proceedings of the 2014 UKSim-
AMSS 16th International Conference on Computer Modelling and Simulation, Cambridge, UK, 26–28 March 2014; pp. 106–112.
[CrossRef]
5. Livieris, I.E.; Kiriakidou, N.; Stavroyiannis, S.; Pintelas, P. An Advanced CNN-LSTM Model for Cryptocurrency Forecasting.
Electronics 2021, 10, 287. [CrossRef]
6. Wirawan, I.M.; Widiyaningtyas, T.; Hasan, M.M. Short Term Prediction on Bitcoin Price Using ARIMA Method. In Proceedings
of the 2019 International Seminar on Application for Technology of Information and Communication (iSemantic), Semarang,
Indonesia, 21–22 September 2019; pp. 260–265. [CrossRef]
7. Lahmiri, S.; Bekiros, S. Cryptocurrency forecasting with deep learning chaotic neural networks. Chaos Solitons Fractals 2019,
118, 35–40. [CrossRef]
8. Adegboruwa, T.I.; Adeshina, S.A.; Boukar, M.M. Time Series Analysis and prediction of bitcoin using Long Short Term Memory
Neural Network. In Proceedings of the 2019 15th International Conference on Electronics, Computer and Computation (ICECCO),
Abuja, Nigeria, 10–12 December 2019; pp. 1–5. [CrossRef]
9. Tandon, S.; Tripathi, S.; Saraswat, P.; Dabas, C. Bitcoin Price Forecasting using LSTM and 10-Fold Cross validation. In Proceedings
of the 2019 International Conference on Signal Processing and Communication (ICSC), Noida, India, 7–9 March 2019; pp. 323–328.
[CrossRef]
10. Hamayel, M.J.; Owda, A.Y. A Novel Cryptocurrency Price Prediction Model Using GRU, LSTM and bi-LSTM Machine Learning
Algorithms. AI 2021, 2, 477–496. [CrossRef]
11. Patel, M.M.; Tanwar, S.; Gupta, R.; Kumar, N. A Deep Learning-based Cryptocurrency Price Prediction Scheme for Financial
Institutions. J. Inf. Secur. Appl. 2020, 55, 102583. [CrossRef]
12. De Gooijer, J.G.; Hyndman, R.J. 25 years of time series forecasting. Int. J. Forecast. 2006, 22, 443–473. [CrossRef]
13. Box, G.E.P.; Jenkins, G.M. Time Series Analysis: Forecasting and Control; Holden-Day: San Francisco, CA, USA, 1976.
14. Yamak, P.T.; Yujian, L.; Gadosey, P.K. A Comparison between ARIMA, LSTM, and GRU for Time Series Forecasting. In
Proceedings of the 2019 2nd International Conference on Algorithms, Computing and Artificial Intelligence, Sanya, China, 20–22
December 2019; Association for Computing Machinery: New York, NY, USA, 2020; pp. 49–55. [CrossRef]
15. Roy, S.; Nanjiba, S.; Chakrabarty, A. Bitcoin Price Forecasting Using Time Series Analysis. In Proceedings of the 2018 21st
International Conference of Computer and Information Technology (ICCIT), Dhaka, Bangladesh, 21–23 December 2018; pp. 1–5.
[CrossRef]
16. Walther, T.; Klein, T.; Bouri, E. Exogenous drivers of Bitcoin and Cryptocurrency volatility – A mixed data sampling approach to
forecasting. J. Int. Financ. Mark. Inst. Money 2019, 63, 101133. [CrossRef]
17. Maciel, L. Cryptocurrencies value-at-risk and expected shortfall: Do regime-switching volatility models improve forecasting? Int.
J. Financ. Econ. 2021, 26, 4840–4855. [CrossRef]
18. Mba, J.C.; Mwambi, S.M.; Pindza, E. A Monte Carlo Approach to Bitcoin Price Prediction with Fractional Ornstein–Uhlenbeck
Lévy Process. Forecasting 2022, 4, 409–419. [CrossRef]
19. Derbentsev, V.; Babenko, V.; Khrustalev, K.; Obruch, H.; Khrustalova, S. Comparative performance of machine learning ensemble
algorithms for forecasting cryptocurrency prices. Int. J. Eng. 2021, 34, 140–148.
20. Chevallier, J.; Guégan, D.; Goutte, S. Is It Possible to Forecast the Price of Bitcoin? Forecasting 2021, 3, 377–420. [CrossRef]
21. Khedr, A.M.; Arif, I.; El-Bannany, M.; Alhashmi, S.M.; ; Sreedharan, M. Cryptocurrency price prediction using traditional
statistical and machine-learning techniques: A survey. Intell. Syst. Account. Financ. Manag. 2021, 28, 3–34. [CrossRef]
22. Brownlee, J. Deep Learning for Time Series Forecasting: Predict the Future with MLPs, CNNs and LSTMs in Python; Machine Learning
Mastery: San Juan, PR, USA, 2018.
23. Andreoni, A.; Postorino, M.N. Time Series Models to Forecast Air Transport Demand: A Study about a Regional Airport. IFAC
Proc. Vol. 2006, 39, 101–106. [CrossRef]
24. Lim, C.; McAleer, M. Time series forecasts of international travel demand for Australia. Tour. Manag. 2002, 23, 389–396. [CrossRef]
25. Lorek, K.S.; Lee Willinger, G. An analysis of the accuracy of long-term earnings predictions. Adv. Account. 2002, 19, 161–175.
[CrossRef]
26. Contreras, J.; Espínola, R.; Nogales, F.; Conejo, A. ARIMA models to predict next-day electricity prices. IEEE Trans. Power Syst.
2003, 18, 1014–1020. [CrossRef]
27. Iqbal, M.; Iqbal, M.; Jaskani, F.; Iqbal, K.; Hassan, A. Time-Series Prediction of Cryptocurrency Market using Machine Learning
Techniques. EAI Endorsed Trans. Creat. Technol. 2021, 8, e4. [CrossRef]
28. Ban, T.; Zhang, R.; Pang, S.; Sarrafzadeh, A.; Inoue, D. Referential kNN regression for financial time series forecasting. Lect. Notes
Comput. Sci. (Incl. Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinform.) 2013, 8226 , 601–608. [CrossRef]
Forecasting 2023, 5 208

29. Troncoso Lora, A.; Santos, J.; Santos, J.; Ramos, J.; Expósito, A. Electricity market price forecasting: Neural networks versus
weighted-distance k Nearest Neighbours. Lect. Notes Comput. Sci. (Incl. Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinform.) 2002,
2453, 321–330.
30. Huang, W. KNN Virtual Currency Price Prediction Model Based on Price Trend Characteristics. In Proceedings of the 2022 IEEE
2nd International Conference on Power, Electronics and Computer Applications (ICPECA), Shenyang, China, 21–23 January 2022;
pp. 537–542. [CrossRef]
31. Vapnik, V. The Nature of Statistical Learning Theory; Springer Science & Business Media: Berlin, Germany, 1995.
32. Wu, S.; Akbarov, A. Support vector regression for warranty claim forecasting. Eur. J. Oper. Res. 2011, 213, 196–204. [CrossRef]
33. Bunescu, R.; Struble, N.; Marling, C.; Shubrook, J.; Schwartz, F. Blood glucose level prediction using physiological models and
support vector regression. In Proceedings of the 2013 12th International Conference on Machine Learning and Applications,
Miami, FL, USA, 4–7 December 2013; Volume 1, pp. 135–140. [CrossRef]
34. Xia, Y.; Liu, Y.; Chen, Z. Support Vector Regression for prediction of stock trend. In Proceedings of the 2013 6th International
Conference on Information Management, Innovation Management and Industrial Engineering, Xi’an, China, 23–24 November
2013; Volume 2, pp. 123–126. [CrossRef]
35. Naing, W.; Htike, Z. Forecasting of monthly temperature variations using random forests. ARPN J. Eng. Appl. Sci. 2015,
10, 10109–10112.
36. Liu, Y.; Sarabi, A.; Zhang, J.; Naghizadeh, P.; Karir, M.; Bailey, M.; Liu, M. Cloudy with a chance of breach: Forecasting
cyber security incidents. In Proceedings of the 24th USENIX Security Symposium, Washington, DC, USA, 12–14 August 2015;
pp. 1009–1024.
37. Zagorecki, A. Prediction of methane outbreaks in coal mines from multivariate time series using random forest. Lect. Notes
Comput. Sci. (Incl. Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinform.) 2015, 9437, 494–500. [CrossRef]
38. Hochreiter, S.; Schmidhuber, J. Long Short-Term Memory. Neural Comput. 1997, 9, 1735–1780. [CrossRef] [PubMed]
39. Ma, X.; Tao, Z.; Wang, Y.; Yu, H.; Wang, Y. Long short-term memory neural network for traffic speed prediction using remote
microwave sensor data. Transp. Res. Part C Emerg. Technol. 2015, 54, 187–197. [CrossRef]
40. Pham, T.; Tran, T.; Phung, D.; Venkatesh, S. Predicting healthcare trajectories from medical records: A deep learning approach.
J. Biomed. Inform. 2017, 69, 218–229. [CrossRef]
41. Solgi, R.; Loáiciga, H.A.; Kram, M. Long short-term memory neural network (LSTM-NN) for aquifer level time series forecasting
using in-situ piezometric observations. J. Hydrol. 2021, 601, 126800. [CrossRef]
42. Wang, B.; Kim, I. Short-term prediction for bike-sharing service using machine learning. Transp. Res. Procedia 2018, 34, 171–178.
[CrossRef]
43. Troia, S.; Alvizu, R.; Zhou, Y.; Maier, G.; Pattavina, A. Deep Learning-Based Traffic Prediction for Network Optimization. In
Proceedings of the 2018 International Conference on Transparent Optical Networks, Bucharest, Romania, 1–5 July 2018. [CrossRef]
44. Becerra-Rico, J.; Aceves-Fernández, M.; Esquivel-Escalante, K.; Pedraza-Ortega, J. Airborne particle pollution predictive model
using Gated Recurrent Unit (GRU) deep neural networks. Earth Sci. Inform. 2020, 13, 821–834. [CrossRef]
45. Muhammad, A.U.; Yahaya, A.S.; Kamal, S.M.; Adam, J.M.; Muhammad, W.I.; Elsafi, A. A Hybrid Deep Stacked LSTM and GRU
for Water Price Prediction. In Proceedings of the 2020 2nd International Conference on Computer and Information Sciences
(ICCIS), Sakaka, Saudi Arabia, 13–15 October 2020; pp. 1–6. [CrossRef]
46. Bai, S.; Kolter, J.Z.; Koltun, V. An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling.
arXiv 2018, arXiv:1803.01271.
47. Zhu, R.; Liao, W.; Wang, Y. Short-term prediction for wind power based on temporal convolutional network. Energy Rep. 2020,
6, 424–429. [CrossRef]
48. Ardimento, P.; Aversano, L.; Bernardi, M.L.; Cimitile, M.; Iammarino, M. Temporal convolutional networks for just-in-time design
smells prediction using fine-grained software metrics. Neurocomputing 2021, 463, 454–471. [CrossRef]
49. Zhang, C.X.; Li, J.; Huang, X.F.; Zhang, J.S.; Huang, H.C. Forecasting stock volatility and value-at-risk based on temporal
convolutional networks. Expert Syst. Appl. 2022, 207, 117951. [CrossRef]
50. Politis, A.; Doka, K.; Koziris, N. Ether Price Prediction Using Advanced Deep Learning Models. In Proceedings of the 2021 IEEE
International Conference on Blockchain and Cryptocurrency (ICBC), Sydney, Australia, 3–6 May 2021; pp. 1–3. [CrossRef]
51. Lim, B.; Arık, S.O.; Loeff, N.; Pfister, T. Temporal Fusion Transformers for interpretable multi-horizon time series forecasting. Int.
J. Forecast. 2021, 37, 1748–1764. [CrossRef]
52. Srivastava, A.; Cano, A. Analysis and forecasting of rivers pH level using deep learning. Prog. Artif. Intell. 2022, 11, 181–191.
[CrossRef]
53. Wang, L.; Mykityshyn, A.; Johnson, C.; Cheng, J. Flight Demand Forecasting with Transformers. arXiv 2021, arXiv:2111.04471.
54. Civitarese, D.S.; Szwarcman, D.; Zadrozny, B.; Watson, C. Extreme Precipitation Seasonal Forecast Using a Transformer Neural
Network. arXiv 2021, arXiv:2107.06846.
55. Shah, I.; Jan, F.; Ali, S. Functional data approach for short-term electricity demand forecasting. Math. Probl. Eng. 2022, 2022,
6709779. [CrossRef]
56. Shah, I.; Iftikhar, H.; Ali, S. Modeling and forecasting electricity demand and prices: A comparison of alternative approaches.
J. Math. 2022, 2022, 3581037. [CrossRef]
Forecasting 2023, 5 209

57. Sopelsa Neto, N.F.; Stefenon, S.F.; Meyer, L.H.; Ovejero, R.G.; Leithardt, V.R.Q. Fault Prediction Based on Leakage Current in
Contaminated Insulators Using Enhanced Time Series Forecasting Models. Sensors 2022, 22, 6121. [CrossRef]
58. Cheng, D.; Yang, F.; Xiang, S.; Liu, J. Financial time series forecasting with multi-modality graph neural network. Pattern Recognit.
2022, 121, 108218. [CrossRef]
59. Hyndman, R.J.; Athanasopoulos, G. Forecasting: Principles and Practice, 2nd ed.; Springer: Berlin, Germany, 2018.
60. Dixit, A.; Jain, S. Effect of stationarity on traditional machine learning models: Time series analysis. In Proceedings of the 2021
Thirteenth International Conference on Contemporary Computing (IC3-2021), Noida, India, 5–7 August 2021; Association for
Computing Machinery: New York, NY, USA, 2021; pp. 303–308. [CrossRef]
61. Dickey, D.A.; Fuller, W.A. Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root. Econometrica 1981,
49, 1057–1072. [CrossRef]
62. Guo, T.; Bifet, A.; Antulov-Fantulin, N. Bitcoin Volatility Forecasting with a Glimpse into Buy and Sell Orders. In Proceedings of
the 2018 IEEE International Conference on Data Mining (ICDM), Singapore, 17–20 November 2018; pp. 989–994. [CrossRef]
63. Chokor, A.; Alfieri, E. Long and short-term impacts of regulation in the cryptocurrency market. Q. Rev. Econ. Financ. 2021,
81, 157–173. [CrossRef]
64. Tanwar, S.; Patel, N.P.; Patel, S.N.; Patel, J.R.; Sharma, G.; Davidson, I.E. Deep Learning-Based Cryptocurrency Price Prediction
Scheme With Inter-Dependent Relations. IEEE Access 2021, 9, 138633–138646. [CrossRef]
65. Song, J.Y.; Chang, W.; Song, J.W. Cluster analysis on the structure of the cryptocurrency market via Bitcoin–Ethereum filtering.
Phys. A Stat. Mech. Its Appl. 2019, 527, 121339. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.