Prediction of Forked Cryptocurrency Prices using Machine Learning Methods

This report explores the prediction of Forked Cryptocurrency Prices in relation to machine learning methods (MLM) using ARIMA, LSTM and Facebook Prophet. It analyzes the close prices of Ethereum Classic and Bitcoin Cash and their correlation with USD and Euro. The report concludes that ARIMA has got a significant accuracy in prediction of forked cryptocurrency. Furthermore, USD showed a strong relationship with the Bitcoin Cash and Ethereum classic on the prediction of Forked Cryptocurrency. The report gives recommendations and areas of further research.

FIN1
Prediction of Forked Cryptocurrency Prices
Student’s name
Tutors name
Course
Institutional affiliation
City and state
Date

FIN2
Abstract
The report explored the prediction of
Forked Cryptocurrency Prices in relation to
machine learning methods (MLM). The
MLM used in this report were ARIMA,
LSTM and Facebook Prophet. The research
questions of the report were; whether
investing in Ethereum classic
cryptocurrency yields the statistical change
in forecasting of forked cryptocurrency.
Also, analyzing the close prices of the
Ethereum Classic and Bitcoin Cash whether
they yield on the volume of trade of
cryptocurrency indexes. The report
concludes that ARIMA has got a significant
accuracy in prediction of forked
cryptocurrency. This was because ARIMA
has got many tests such as PACF, ACF that
show clearly the correlation and the
stationarity of lagged values for prediction.
Furthermore, USD showed a strong
relationship with the Bitcoin Cash and
Ethereum classic on the prediction of Forked
Cryptocurrency. Also, the report found out
that the data about variables in the report
were normally distributed. The report further
gave recommendations and areas of further
research.
1.0 Introduction
Bitcoin Cash
Bitcoin Cash is considered to
be an improved component of the
bitcoin core software. It was

FIN3
established in 2017 August, 1st. After
the introduction of bitcoin cash, the
block size was seen to increase to
8mb. From then, the processing of
more payments per second by miners
has been observed (Karif, 2019). The
transactions of cash have become
cheaper and faster and this makes it
easier for the users. The main cause
of separating bitcoin cash from
bitcoin in August was the much heat
and persistent debate concerning the
digital currency all over the country.
However, after some little time, the
debate went on till November when
an announcement arose indicating
another improvement of the bitcoin
software (Kuzmina, 2019. After a
minimum of six months since the
announcement was made, questions
as to why there is need to make a
software upgrade immediately. The
Coin desk wrote a report answering
the questions. According to their
report, they explained that the new
bitcoin cash software would help the
processing network to make more
transactions compared to the existing
bitcoin. In addition, bitcoin itself had
become limited and barbaric in its
usage. Currently, the bitcoin cash is
progressing faster according to its
founders compared to the former
bitcoin. However, the developers of
bitcoin cash resisted to take on the
ideas of bitcoin founders that the
block size that was increased, could
lead to malfunctioning of the
network.
Ethereum Classic
In May of the year 2016, the
capital venture fund that was termed
as the DAO developed on Ethereum
initiated with around $168 million.
This was done with the aim of
investing in profitable ventures with

FIN4
the aid of smart contracts. Most
significant if that, in June 2016 about
50 million USD was got from the
accounts in the DAO and transferred
on another account whereby the
owner’s consent was not put into
consideration (Andreas et al, 2018).
However, this exploited some of the
vulnerabilities which were
introduced in the May of 2016. This
made the members of the DAO and
the community of Ethereum to meet
and debate about the actions to
resolve such incidence (Megas.
2018). The Vote was casted in July of
2016 where they implemented the
usage of hard fork in form of
Ethereum code. The ideology of
developing hard fork as a separate
Cryptocurrency was mostly on the
officials’ minds after some weeks
from the incidence of hackers.
Different parties had not acquainted
themselves to get some time to
discuss about the opportunity of
Ethereum channel that is official as
at first the motion was forbidden.
Following only the denials, still the
concept of developing the Ethereum
Classic grew (Kuzmina, 2019). As it is
observed that it was first the project
being managed by the entire
community, the community
officiated the logo, website content
and the name. Therefore, Ethereum
Classic came into existence after the
website launch since the token was
summed up to the exchange of
Poloniex (Lucey et al, 2018).
The existence of the
Ethereum Classic came out after the
members of Ethereum community
had refuted the hard fork about the
immutability grounds. According to
the principal of the blockchain, the
hard fork cannot be changed and

FIN5
adopt the usage of version of
Ethereum. that is unforked.
Moreover, the Ethereum Classic that
came at first was involved in the
forked Ethereum chain with the
number of block of 1,920,000
(Andreas et al, 2018). This was created
by the Classic miners in 2016 in the
month of July. Also, Ethereum
Classic went through the hard fork
technical in order to alter the pricing
internally for different Op codes
especially on the Ethereum Virtual
Machine (EVM). It is evident that
individuals who followed Ethereum
Classic takes take on the
immutability of blockchain against
Ethereum coins.
Ethereum Classic background
is among the things whereby one
cannot just tame and understand it.
The details about Ethereum Classic
seems not to be gathered from one
place, this makes most of the people
who are newly on the space to
understand it (Phyro, 2018). This is
because the DAO hard fork did not
respond on reverting the transactions
got from the Hacker. In addition, the
account was vanished with the
balances estimated 100 wallets and
instead yielded other wallet balances.
However, with all these situations,
Ethereum Classic has been made as
vital rule that is used as individual
bank after fighting every kind of
obstacles especially with hard fork at
the beginning and the entire
community (Lerer, 2019).
Ethereum is said to be among
the most prominent Crypto currency
all over the world. The major aim of
creating these coins was to control
the problems and limitations of
Bitcoin Cash and extend the block
chain technology boundaries (Li et

FIN6
al, 2017). With Ethereum Classic,
one is able to produce any kind of
applications on its blockchain one
thing that Bitcoin Cash can never do
(Hayes, 2018). Besides this,
someone who wished to create a
block-chain-based application to be
able to create their blockchain from
scratch. With the use of Ethereum,
the creators could create any
imaginable application through
leveraging off the Ethereum
infrastructure. Ether (ETH) is the
name given to the native currency of
the Ethereum block chain. Gas is the
cost computation of Ethereum
blockchain.
The two major aims of the
Ethereum Classic coins were;
accepting the idea that hack
happened and that since the
blockchain behaves like an
immutable ledger, nothing could be
done. Also there was need to
implement the Ethereum classics a
tool for stopping transactions and
hackers from taking the money
(Kuster, 2017). The current
Ethereum classic of the Ethereum
protocol is Ether Zero.
1.2 Motivation of the report
Fluctuations in the Forked
Cryptocurrency Prices; with
cryptocurrency, there are several
occurrences of fluctuations in the
prices and the general number of
transactions of Forked
cryptocurrency. For example, with
the most popular cryptocurrency,
Bitcoin Cash and Ethereum Classic
had already witnessed that there are
no significant changes in their prices
and in the number of transactions
especially in the year of 2018.
When the Bitcoin Cash

FIN7
started to roam about the world’s
market intention, this witnessed that
there was a significant rise and some
fluctuations among the prices and the
number of times they are transacted.
In relation with other
cryptocurrencies, Litecoin and ripple
forms, have been with unstable
changes since December of 2016
(Luther, 2016). These unstable prices
and unstable transactions have given
various opportunities about the
speculation about users in hindering
other possible users to invest in
Forked cryptocurrency (Alvarez-
Ramirez et al, 2018). High rates of
risks being assumed by the people;
most of the users always have bias
towards Bitcoin Cash and Ethereum
Classic. Due to the rise of
competition being attributed from the
side of other means of money such
as banks and other financial
institutions, these institutions have
been rapidly shifted to the
deliverance of the business
operations. Also the financial
Technologies (Fin Tech) and the
banks on the national levels have
been seen competing with some
financial groups such as Bitcoin
Cash and Ethereum Classic (Luther
& Salter, 2017). These banks are
working on the Ethereum Classic to
increase the process of the
businesses by liaising the endurance
of customers to enter in the
stipulated competition among the
Forked Cryptocurrency prices and
transactions (Alvarez-Ramirez et al,
2018). There is the emergence of the
casual relationship which is between
the knowledge of customer
management. This relationship can
bring some biases about the risks and
changes or fluctuations in the prices

FIN8
of such businesses. Such challenges
have disallowed the existing cycle of
the customers to engage fully
cryptocurrencies (Luther & Salter,
2017). The bias of the business
among the customers can be resolved
by increasing the loyalty and churn
reduction so as to make them engage
fully in the business.
Emergence of haphazard
regulations; several studies note that
there is regulation of coherent in
terms of inhibiting the adoption and
building trust in Bitcoin Cash and
Ethereum Classic coins as per
Forked Cryptocurrency is concerned
(Thies & Molnar, 2018). It is evident
that for something to be called
money as in crypto terms, it must full
fill three major values. These values
include the following; medium of
exchange, value of account, and the
value of storage. With all these core
values, regulations come in since
they generate income to individuals
and the governments pf the countries
that engage in Bitcoin Cash and
Ethereum Classic (Alvarez-Ramirez
et al, 2018).
The crypto space has gone
on the scam of the wide spread, the
consistent usage of technology is
also in illegal functions and which
are likely to be not put to an end
especially in supply themed drug
cryptocurrencies (Tiwari et al, 2018).
However, the reputation of Ethereum
Classic and Bitcoin Cash as per
Forked Cryptocurrencies is
concerned the legitimate proposed
projects and giving out the sorry and
alarming state of signals to those
who adopt to the change with the
pricing of forked cryptocurrency.
The solution which is natural has got

FIN9
some regulation problems, thus, the
nature of global financial
technologies which impact to the
attributor of problems about the
measures of traditional regulations.
Also, some studies about the global
cryptocurrencies are in the state of
haphazard and the way of piecemeal
situations (Anne- HauboDyhrberg,
2016). This is due to other countries
about Bitcoin Cash and Ethereum
Classic have been committed in
operating together on the regulations
while other nations are operating the
regulations of the cryptocurrencies in
the opposite ways.
The possibility of volatile and
vagueness of the immense collapse;
the process of volatility is likely to
be put off from the features of the
value of storage. The coins which
stable attribute to proper working
situation in Forked Cryptocurrency
(Maesa et al, 2017). There is the
emergence of problems of the
cryptocurrency as assumed to be out
righting to zero with the number of
risks in routing funds by using the
centralized exchange rates.
1.3 Context of the report
The prediction of Forked
Cryptocurrency prices of Bitcoin
Cash and Ethereum classics in
estimation of Forked crypto
currencies using the Machine
learning methods (MLM) can be
clearly indicated by either
statisticians or any other personnel
with related field (Aru, 2019).
Unstable prices and unstable
transactions have given various
opportunities about the speculation
about users in hindering other
possible users to invest in
cryptocurrencies. This has been seen
from several researches that, studies

FIN10
about the attributes of the currencies
which are likely to progress all over
the world (Lahmiri &Bekiros, 2018).
In addition, most studies find that the
sentiments of the users in relation to
crypto currencies on media as well as
the changes in the number of trade
and prices to yield any relationship.
Most of the reliable studies have
been zeroed on the Bitcoin Cash
since the big amounts of data that it
gives in order to eliminate the value
of building the model to forecast
fluctuations in the number of
transactions and prices of
cryptocurrencies (Maesa et al, 2017).
1.4 Research Purpose
The report aimed at analyzing
the prediction of Forked
Cryptocurrency prices with the aid of
machine learning methods. This
report is hoped to predict the closing
prices of Ethereum Classic and
Bitcoin Cash in relation to the
machine learning algorithms. Also
the report utilized several other
specific objectives and these
included;
 To determine the model with
efficient and accuracy prediction of
closing prices about Ethereum
Classic
 Determine the strength of BCH and
ETC in correlation with USD and
Euro
 Determine the comparative
evaluation between Bitcoin Cash
and Ethereum Classic.
1.5 Research Question
 Which approach or model performs
the best outcome in form of
efficiency and accuracy prediction of
closing prices about Ethereum
Classic?

FIN11
 How strongly or weakly are ETC and
BCH in correlation with USD and
Euro?
 What is the comparative analysis
between the Bitcoin Cash and
Ethereum Classic?
1.6 Impact and Implication
of the Report
The report is hoped to
provide the current literature about
the pricing of Bitcoin Cash and
Ethereum classic in prediction of
Forked Cryptocurrency Prices. More
so, the report hopes to provide the
conclusions and findings to the
legislation authorities, governments
of the nations and other bodies
concerned with digital currency and
the fin techs (Urquhart, 2016).
Related literature
Shah (2017), carried out a
report on the Bayesian regression
model in generating models of the
latent sources. The datasets of
Bitcoin Cash were taken from Okoin
found in China. His report focused
on collecting the relevant data about
the prediction of the price in every
10minutes in usage of exchange of
Bitcoin Cash cryptocurrency. The
findings of 50 days showed ROI was
89% with the ratio of Sharp at 4.10
(Shah, 2017)
In addition, McNally in
2018conducted a report about the
several modeling experiments on the
long short-term memory (LSTM).
Autoregressive Integrated Moving
Average (ARIMA) and Recurrent
Neutral Network (RNN). According
to his report, the approaches were
generated in accordance to the Open,
High and low close data emanating
from Bitcoin Cash and the data from

FIN12
Hash rate in Block chain. The report
found out that the highest accuracy
was at 52.78% with its Root Square
Mean Error of 5.45%. (McNally,
2018)
Nikolaos and Kyriazis (2019)
conducted a research on efficiency
and profitable trading opportunities
about Forked Cryptocurrency
markets. The report used a
systematic survey to determine
whether the pricing behavior of
Forked Cryptocurrencies can be
forecasted. Therefore, the efficient
market hypothesis was not accepted
and meditating was practicable via
trading. Methodologies on testing
long memory in returns and volatility
that were used by Nikolas included
the Rescaled Range (R/S), De-
trended Fluctuation Analysis (DFA)
and others (Nikolaos &Kyriazis,
2019). The researcher discovered
that many academic papers give
evidence for inefficiency of Bitcoin
Cash and several other digital
currencies that are very important.
However, to achieve various steps
have been gone through for the
previous years. They concluded that
speculators can have lower trading
strategies that are profitable.
Andrew (2018) carried out a
research on Exploring the
Interconnectedness of Crypto
currencies using Correlation
Networks. The researcher detected
characteristics which are fixed over
time influence the price evolution
over time positively using correlation
networks. The websites and white
papers of Crypto currencies having a
larger number of user-bases were
used to identify potentially important
characteristics. Two datasets were

FIN13
used to asses these characteristics to
enhance robustness. The first had
fourteen Crypto currencies starting
from November 2017 and a subset
had nine Crypto currencies
beginning from September 2016.
Both these were to end in March
2018. Mere analyzing only the subset
of Crypto currencies made the data
points to raise from 115 to 537. This
raised the changes of robustness in
relationships over a given time
period. Minus the USD Tether, the
results from the findings showed a
positive relationship between the
different types of Crypto currencies
and they were statistically significant
(Koshy et al, 2014). Besides, stronger
positive associations were seen for
six Crypto currencies. One was the
Ethereum classic and the other was
the Bitcoin Cash which was
exceptional. The researcher
evidenced the existence of a many
Crypto currencies majorly associated
with cardano and the other group
correlated with Ethereum (Lagoarde-
Segot, 2015) The results were not
consistent with the functionality of
the token and the mechanism of
creation was discovered to be the
most dominant determinant for the
price evolution over the period.
However, the results did not provide
the factors that contributed to the
price speculated (Kuster, 2017).
Mira (2018) conducted a
research on “An Analysis of
Cryptocurrency Governance”. The
researcher found out that many
people have been excited with the
promise of reducing governance
through the use of digital currency
(Crypto currency) to make better
usage of block chain technology and
cryptography to carry out

FIN14
transactions. Crypto currency
depends on large relationships
between various parties with
different functions and
responsibilities though the structure
of Crypto currency is decentralized.
The effect of Cryptocurrency
therefore relies on the mutually
satisfying occurrences of these
parties who are the base for non-
technical governance structures.
Mira therefore investigated the
extent to which technical governance
reduces traditional governance
challenges through examining the
governance structure of the two
types of Crypto currencies that is; the
Bitcoin Cash and Ethereum coins.
Firstly, it provides the background
for the origin and technical value
consideration of Crypto currency and
also the governance theory before
analyzing Bitcoin Cash and
Ethereum to investigate whether
technology reduces the party’s
motivations. The findings of Mira
show that both Bitcoin Cash and
Ethereum depend mainly on trust
networks showing the elements of
non-technical governance are
important in their effectiveness
despite of the promise of reducing
governance (Mira, 2018).
Chapter three:
Methodology
3.0 Introduction
The chapter consists of the
extracted data, the research design of
the report, efficiency and accuracy of
the currency hypothesis, the
techniques of analyzing data. The
techniques of analysis involved

FIN15
stationarity, linearity and the
normality tests among the variables
in the report in relation to the nature
of the data used.
3.1 Research design of the
Report
The research adopted two
designs. These designs included the
descriptive and correlational designs
(Robert, 2018). The reason of using
these kind of research designs was
the report aimed at acquiring the
positivism kind of research. In the
positivism context, the concentration
is put on the quantitative data from
the trusted sources of data. This
therefore helped the researcher to
obtain current information about
existing literature and also predicts
the variables in the report. Most
significantly, the data used in the
report was secondary based.
3.2 Data
In this report, the on-chain
volume of data was used. These data
include the following; BTC, BCH
among others. The data was
extracted from the block chain which
are aggregated by the tools of python
and Bitcoin Cash tokens (Urquhart,
A. 2018). Basing on some fields, the
TX Volume (USD) influences the
pricing of cryptocurrencies around
the globe. It is therefore wider and
broad to measure the degree of
outcomes of the block chain and
cryptocurrency at large. The main
interest here is how critical the value
of the USD denomination can
circulate in terms of classic coins and
Bitcoin Cash on the day to day block
chain.
Also, the close pricing of the
classic coins and Bitcoin Cash
currencies can be measured in terms
of transaction via online (Delgado-

FIN16
Segura et al, 2018). This can be such
a difficult thing to be estimated since
day and night the values and the
figures about the currency change
and fluctuates at any time (Valenzula
,2017). But however, the improper
estimates of the pricing about the
coins can be put in its way of
funding to produce good results and
the closer block-chain of pricing to
yield Forked Cryptocurrency.
Furthermore, the TX Count data
helped in getting the analytical
perspectives in this report. This kind
of data deals with the transactions
which are in operation of the block
chain on the public note on the
specified time. It must be put on the
significant note that, the lowness of
block chain and the Bitcoin Cash
fees can easily be fabricated as the
total bunch as transactions are
concerned (Turk &Klinc, 2017). It is
evident that TX Count provides the
data about the Ethereum Classic and
Bitcoin Cash data which is likely to
be underestimated in prediction of
forked cryptocurrency.
Testing Efficiency of
Currency Hypothesis about
Ethereum Classic and Bitcoin
Cash
In testing the currency
hypothesis in the digital market, the
framework put forward by the
pioneer of hypothesis concerning the
currency efficiency. The pioneer of
efficiency of currency hypothesis
was developed using the ideas of
Nobel Laurite Egene in the year of
1965 (Vandezande, 2017). The
changes in prices are considered to
be the independent successive points
of the close pricing of Bitcoin Cash
and Ethereum Classic coins (Liu,

FIN17
2018). In this report, parametric tests
will be used in moist cases since the
report aimed at achieving the current
information (Vidal-Tomás & Ibañez,
2018). These tests are the
Augmented Dickey Fuller tests (Unit
root test) among others.
3. 3 Methods of Data
Analysis and Tests
The analysis under this report
was majorly retrieved from the
online sources of Bitcoin Cash and
classic coins in relation to the index
of crypto currencies. Also, data was
analyzed using Statistical Package
for Social Scientists (SPSS) and
STATA statistical packages to create
the normality tests and linearity tests.
The linearity tests were done using
the regression analysis.
Augmented Dickey Fuller test
Augmented Dickey Fuller test
(ADF) is used in testing the analysis of
stationarity (Unit root test). This shows
whether the variables in the report are
attributed to the sequence or trend of the
pricing of Bitcoin Cash and classic coins
in relation to the index of
cryptocurrencies (Lahmiri&Bekiros,
2018). Under the Augmented Dickey
Fuller test, there is the possibility of
testing about the null hypothesis in
determining whether the unit root test
exists. It is evident that when there is the
possibility of unit root, the sequence
about the data becomes random walk
(Vidal-Tomás&Ibañez, 2018).
Therefore, such sequence depicts the
trending data about the variables in the
report. Also, such test can be achieved
from mean, median and variances which
have got values which are constant. The
data about the Augmented Dickey fuller
test using auto-regression method with
no intercepts about the trend is state as

FIN18
below:
Normality Tests
Basing on this research, the
tests of normality about the variables
under the report were tested using
the P-P plots, and Shapiro Wilkson
tests. The criterion is when the
statistic test in Shapiro test is
approximately to 1 (Lucey et al,
2018). This therefore means that the
data about the variable is said to be
originating from the population
which is distributed normally.
Vector Error Correction
Model (VECM)
Vector error correction model
gives the offer to use vector
autoregressive model which is
known as VAR, this can be done
using the time series data. The
important significant issue here is the
regression which is the student’s test
(t-statistic). The t-statistic is guided
by the significance of R-square, R-
square involves the degree how the
independent variable impacts on the
dependent variable variations
(Lahmiri&Bekiros, 2018). Also
when the R-square value is greater
than 0.6, it shows that there is a
significant relationship between the
variables in the report (close pricing
of Bitcoin Cash, Ethereum classic
and Forked Cryptocurrency. The
importance of conducting VECM
involves steps such as; specifying
and estimation of the VAR model,
calculation of the likelihood tests and
determination of the number of co-
integration.
Time Series data; kind of
data is a series of numbers that is
considered as long side time
intervals. The LSTM and ARIMA

FIN19
models used in prediction act as the
algorithm supervised under the
version of auto-encoder. Such kind
of datasets are split into the outputs
and outputs. However, LSTM model
has a great comparison with the
statistics that is classic in linear
models. This is because it handles
forecasting problems with multiple
inputs. In this report, LSTM used
data to forecast the 30 days in front
of the closing prices (Lucey et al,
2018). The output dataset bases on
the window size that handles the size
of prediction especially the 30 days.
Furthermore, Bitcoin Cash and
Ethereum coins being considered as
the valuable currencies all over the
world, it is better for the ARIMA
model to be trained in order to
experience price predictions. Since
ARIMA can be used for the long
term period predictions, the
emergence of large error predictions
is likely to occur in foresting of price
behavior (Alsindi, 2019).
For the case of Prophet
model, it looked at the quickest
means of retrieving the prediction of
the Bitcoin Cash and Ethereum
coins. This is because it gave the
open source of extraction of data
about the time series (Zou, 2018).
The data about the prophet
approach can be got from the python
and R-languages. Using the times
series data, RMSE is got from
running the regression analysis
between the variables in the report.
This helps to identify the degree of
difference between values predicted
by a model.
Also, in most times series
data, the Mean Absolute Percentage
error is used in determining the
accuracy of the system of prediction.

FIN20
It can be determined by the usage of
percentages of absolute values in
each time period.
M = 1
n ∑
t =1
n
¿ At −Ft
At where At
is the actual time and F is the
forecast time. However, with MAPE,
extreme values are only taken
without the consideration of negative
values.
Overview of machine Learning models
LMSTM
Long short-term memory
(LSTM) is referred to as the Man-
Made recurrent neutral network
(RNN) architecture that is vital in the
deep learning fields. In contrast with
the neutral networks that are
standard feedforward, LSTM has got
the connections of feedback that alter
it into the ‘general purpose
computer’ (implying it computes
anything z machine that is Turing
can do) (Haider et al, 2019). The
output of the LSTM can be
modulated by the cells. They are
used in predicting sequence
problems especially in complex ones
like machine translation and
recognition of the speech.
PROPHET
Prophet which is at times
called the Facebook prophet model is
a procedure in forecasting the time
series data. This can be done basing
on the additive model in that there
are non-linear trends to be fitted. The
time series data can be categorized
hourly, weekly, monthly, quarterly
and yearly (Taylor & Letham, 2017). It
is hard for many machine learning
methods to give high quality
predictions since it requires
substantial requirement of
experience and ‘specific skills’. This
method allows both the non-experts
and the experts in the field of

FIN21
forecast the quality.
ARIMA model
The ARIMA model is a short form of
Autoregressive Integrated Moving Average
Model. It is considered under statistical
models that are applied in analyzing and
forecasting data in time series. The model
majorly involves a set of standard structures
within time series data. It is very useful and
applicable while forecasting a strong time
series (Juang et al, 2017).
In time series data, Auto-correlation
Function (ACF) shows the relationship of
the lagged values. ACF provides the
researchers with the values about
autocorrelation about the sequence on
lagged values (Alsharif et al, 2019). The ACF
is plotted with the corresponding values of
confidence band. Similarly, it describes how
the series of lagged values that are present
with those in the past.
Furthermore, Partial Auto-correlation
Function (PACF) in prediction criteria
takes the consideration of relationship
between ‘intermediated lagged values’ and
the time series data. It further shows the
stationarity in the time series data (Alsharif et
al, 2019) When testing the stationarity, this
helps to know whether data about the
variable say Forked cryptocurrency is
trending or not.
In univariate where there is a single vector,
ARIMA is applied as a technique for
forecasting future values of self-relying
series. The model is used majorly in
situations where there is short term
prediction that needs more than 40 points
past data. Learning algorithms is the term
given to the most suitable time series
analysis required in forecasting like
Exponential smoothing or the ARIMA
models. These models are said to be
examples of machine language (ML) used
for regression (Juang et al, 2017).
Results
The report aimed at exploring

FIN22
the impacts of pricing of Bitcoin
Cash and Ethereum Classic using the
Machine Learning Method in
prediction of Forked Cryptocurrency.
However, there are other specific
objectives under which the
researcher intends to investigate on.
These objectives included the
following: Investigate whether
Ethereum Classic Cryptocurrency
can yield the significant change in
form of roaming investment in
currencies. And Analyze the closing
price of Ethereum Classic prices on
how to trade the volumes of the
increase of forked cryptocurrency
(Vranken, 2017).
From 2017 to December
2018, shows that there was a
decrease in the number of
transactions of Bitcoin Cash. From
2017 up to current that is 2019, there
have been a steady increase of
Bitcoin Cash transactions using the
logarithmic scale (Lucey et al, 2018).
This showed that Forked
cryptocurrency is increasing since
the public among the nations are
gradually adopting it. This situation
was achieved not only concentrating
on the ideas of communities of
cypherpunk but also relying on the
authorities of the state. More so, the
performance of cryptocurrency has
undergone the high rate of growth in
terms of significant currency inform
of online and offline. This was due
from 2017 since some businesses
began adopting the usage of Bitcoin
Cash as the tradition currencies.
Furthermore, Ethereum
Classic is not only the payment but
also the payment system. The
founder of Ethereum believed that
the block chain has impacted more
utilization as just been the service

FIN23
provider system. The contract that is
considered to be smart was coded
running on top of the block involving
a set of laws and regulations through
which the participants of the agreed
contracts interact with one another
(Anderson et al, 2016).
Furthermore, there is a basis
of price volumes (Bitcoin Cash and
Ethereum classic coins) to forecast
the Forked cryptocurrency prices.
The scores of the error shows that
there is a decreased rate at a drastic
rate. However, the networks do not
forecast actual values but the
estimates of pricings are usually
made. Also, this incidence returns
the price and the value of the coins to
predict the data about the Forked
Cryptocurrency market.
Most significantly, the report
depicted the pricing of Ethereum
classic and Bitcoin Cash as the
correlates of cryptocurrency index as
per digital money is concerned. In
2017, the prices of the digital
currency that is Bitcoin Cash and
Ethereum classic coins were
relatively changing from month to
month. This meant that there are
several factors which affect the
growth and the performance of
cryptocurrency across the globe. It is
evident that, the pricing of the
Bitcoin Cash and Ethereum classic is
cyclical in nature since it changes
day and night (Vigne et al, 2017).
The analysis of this report
using ARIMA models, the close
pricing of Bitcoin Cash and
Ethereum Classic coins were
significant. This was because the Sig
value of Bitcoin Cash pricing while
using the time series data. This
meant that the value of pricing about
the Ethereum Classic coins and

FIN24
Bitcoin Cash greatly and statistically
influence the prediction of Forked
Cryptocurrency prices of around the
globe in the nations where it
operates.
Conclusions
From the findings above, ETC has been in
stalling for long with $10 marks. However,
the company leaves nothing un changed to
engage in the present and current members
in the community on the platform.
This impacted the adoption of the taken
Ethereum Classic that may sudden hint up
across $10 mark in the coming period. This
situation could be regarded as the better
period for daily trading with the long term
investment as per ETC coin is concerned
due to its movements which show stability
and the credibility in the future.
The report found out that Bitcoin Cash,
Ethereum classic and Ethereum classic
cryptocurrencies as a whole are being
widely used an are growing at a faster rate.
It is evident that they both have big slump
over the past years especially in 2017 and
2018 when people had a lot of bias about the
digital currencies (Blockchain journal,
2019).
With LSTM, an extension of
the networks especially the classic
recurrent addressed the problem of
vanishing gradient in both currencies
of USD and Euro. This is because
the slope or gradient approaches to
zero when the error changes with
different layers of the currency.
Moreover, the long-short term
memory (LSTM) was found of using
the input-forget and the gate of
output which is a bit confusing to
draw conclusions. However, the
trend showed that the final market of
capitalization about the Ethereum
Classic Cryptocurrency in 2017 was
215 billion in USD not like in

FIN25
November of 2016 with 800 billion
USD.
By using the Facebook
Prophet model, the end results were
obtained but simply it was attributed
to messes using this model.
Ethereum Classic and Bitcoin Cash
using the blue fit in the analysis part,
it showed the slight changes in fitting
the Ethereum classic
cryptocurrencies. With this kind of
machine learning method, some
values do not depict exert what the
digital currencies do (Alessandretti et
al, 2018).
By considering the ARIMA
models, this was used in finding the
statistics about time series data. Time
series data helped the researcher to
find the stationarity and linearity
about the variables in the stud.
Generally, ARIMA model takes the
biggest percentage in predicting the
Forked cryptocurrency prices.
Most significantly, USD
prices of Bitcoin Cash and Ethereum
classic was analyzed. However, USD
showed a greater impact in the
variation of prices of the currencies.
This was evident that the machine
learning methods in this report
showed a cyclical movement in
USD, thus predicting or forecasting
higher returns of Bitcoin Cash and
Ethereum Classic
According to the findings of
this report, it is evident that there is
always the low forecasting rate or

FIN26
score as per the Stanford trading of
Bitcoin Cash is concerned. The
report further found out that the
block chain data extracted from
Bitcoin Cash showed the distribution
of close pricings that are spatial in
nature.
Also, the findings showed
that there is high prediction rate of
cryptocurrencies indexes over the
predicted period. This was evidenced
from the cyclical upward movements
of the prices of Bitcoin Cash and
Ethereum Classic.
Most significantly, the
researcher recommends the central
authority of digital currencies and
the governments of the nations
involved in Bitcoin Cash and
Ethereum Classic to always put
regulations which favor the business
of online trading.
Also, the researcher
recommends the cryptocurrency
agencies of Bitcoin Cash, Ethereum
Classic coins, Fintegri, among others
to always increase their value of
close pricing. This increase the
possibility of increase in the value of
cryptocurrencies.
Also, the governments and
authorities of the Bitcoin Cash and
other online or digital currency
traders should increase awareness of
the cryptocurrencies. This is because
most of the people have bias on the
online trading because of ignorance.
Areas for further Research
Basing on the results of this
report, the researcher recommends
other researcher to conduct several
studies on the effect of attitude and
academic levels towards the
operation of indexes of
cryptocurrency. This is because the

FIN27
report showed that there are other
factors affecting the index of
cryptocurrency other than the close
pricing of Bitcoin Cash and the
Ethereum Classic coins.
In addition, the researcher
recommends other researchers to
carry out studies on the effect of
government regulations and policies
put in governance of digital currency
trading.
References
Andrew, Y. 2018. Cryptocurrency
correlation. Retrieved from:
https://www.coindesk.com/2020-candidate-
andrew-yang-pledges-to-create-clear-
guidelines-for-crypto-assets
Antonio. J.N, Mario. I.C , Carlos. A.F.
2019.Statistical analysis of Bitcoin Cash
during explosive behavior periods.
Retrieved
fromhttps://doi.org/10.1371/journal.pone.02
13919.
Anne -HauboDyhrberg. 2016. Bitcoin
Cash, gold and the dollar – A GARCH
volatility analysis[J]. Finance Research
Letters,16 85–92.
Alessandretti. L, Abeer. E, Aiello. L.M and
Baronchelli. A, 2018. Anticipating
Cryptocurrency Prices Using Machine

FIN28
Learning. Complexity Volume 2018, Article
ID 8983590, 16 pages
https://doi.org/10.1155/2018/8983590
Alsharif, A.M., Younes, K.M. & Kim, J.
2019. Time Series ARIMA Model for
Prediction of Daily and Monthly Average
Global Solar Radiation: The Case Study of
Seoul, South Korea. Retrieved from:
https://www.mdpi.com/2073-8994/11/2/240
Aru. L, 2019. How to Mine Ethereum
Classic, Step by Step (with Pics). Bitcoin
Market Journal. Available at.
https://www.bitcoinmarketjournal.com/mine
-ethereum-classic/
Blockchain journal, 2019. The developer
proposed to make Ethereum Classic a base
chain for Ethereum. Available at:
https://blockchainjournal.news/the-
developer-proposed-to-make-ethereum-
classic-a-base-chain-for-ethereum/
Alsindi. W, 2019. Ethereum Classic: The
Ungoverned Blockchain?
https://medium.com/coinmonks/ethereum-
classic-the-ungoverned-blockchain-
b9ae8986a60a
Andreas M. Antonopoulos, Gavin Wood
(2018): "Ethereum Timeline" . In Mastering
Ethereum: Building Smart Contracts and
DApps, O'Reilly Media; pp. 329-424 pages.
ISBN 9781491971918
Anderson.L, Holz. R, Ponomarev. A,
Rimba.P, and Weber. I. (2016). New kids on
the block: an analysis of modern
blockchains. arXiv preprint
arXiv:1606.06530
Croman.K, Decker.C, Eyal. L,.Gencer.A.E,
Juels.A, Kosba.A, Miller.A , Saxena.A,
Shi .E,. Sirer.E.G, et al. 2016. On scaling
decentralized blockchains. In International
Conference on Financial Cryptography and
Data Security, pages 106–125
Alvarez-Ramirez, J., Rodriguez, E., &
Ibarra-Valdez, C. 2018. Long-range
correlations and asymmetry in the Bitcoin
Cash market. Physica A: Statistical

FIN29
Mechanics and its Applications, 492, 948–
955.
Ardia, D., Bluteau, K., &Ruede, M. 2018.
Regime changes in Bitcoin Cashgarch
volatility dynamics. Finance Research
Letters (in press). Atzori, M. 2015.
Blockchain technology and decentralized
governance: Is the state still necessary
Balcilar, M., Bouri, E., Gupta, R.,
&Roubaud, D. 2017. Can volume predict
Bitcoin Cash returns and volatility? A
quantiles-based approach. Economic
Modelling, 64, 74–81
Bariviera, A. F., Basgall, M. J., Hasperué,
W., &Naiouf, M. 2017. Some stylized facts
of the Bitcoin Cash market. Physica A:
Statistical Mechanics and its Applications,
484, 82–90.
Baur, D. G., Dimpfl, T., &Kuck, K. 2018.
Bitcoin Cash, gold and the US dollar-A
replication and extension. Finance Research
Letters, 25, 103–110.
Delgado-Segura, S., Pérez-Sola, C.,
Navarro-Arribas, G., & Herrera-
Joancomartí, J. 2018. A fair protocol for
data trading based on Bitcoin Cash
transactions. Future Generation Computer
Systems, 26, 145–149.
Demir, E., Gozgor, G., Lau, C. K. M.,
&Vigne, S. A. 2018. Does economic policy
uncertainty predict the Bitcoin Cash
returns? An empirical investigation. Finance
Research Letters.
Hayes. A. 2018. Bitcoin Cash price and its
marginal cost of production: supporting
evidence. Applied Economics Letters.
Retrieved
from:https://www.researchgate.net/publicati
on/317601872_Bitcoin
Cash_price_and_its_marginal_cost_of_prod
uction_supporting_evidence
Haider , S. A., Naqvi S.R , Akram, T.
Umar, A.G. , Shahzad, A. , Sial, M.R.,
Khaliq, S. & Kamran, M. 2019. LSTM
Neural Network Based Forecasting Model
for Wheat Production in Pakistan. Retrieved
from: https://www.mdpi.com/2073-
4395/9/2/72
Haseeb. Q, 2017. Why Bitcoin Cash is Not

FIN30
Trustless. Retrieved from:
https://hackernoon.com/Bitcoin Cash-is-not-
trustless-350ba0060fc9. 17)
Juang, W. Huang, S. Huang, F. , Cheng,, P
& Wann, S. 2017. Application of time
series analysis in modelling and forecasting
emergency department visits in a medical
centre in Southern Taiwan. Retrieved from:
https://www.ncbi.nlm.nih.gov/pmc/articles/P
MC5719313/
Kuzmina, C. 2019. The Painful History of
Ethereum Classic. Retrieved from:
https://news.bitstarz.com/the-painful-
history-of-ethereum-classic
Kharif, O (2019)."Ethereum Classic
Movements Halted by Coinbase on Signs of
Attack". Bloomberg. Retrieved from:
https://en.wikipedia.org/wiki/Ethereum_Clas
sicKuster. F. 2017. The War of Crypto
currencies: Ripple vs. Etherium vs. Bitcoin
Cash. Retrieved from:
https://captainaltcoin.com/ripple-vs-
ethereum-vs-Bitcoin Cash/ (describing
Bitcoin Cash as “frictionless, anonymous,
and cryptographically astonishingly secure.”
Lahmiri, S., &Bekiros, S. 2018.
Chaos,randomness and multi-fractality in
Bitcoin Cash market. Chaos, Solitons &
Fractals, 106, 28–34.
Li, X., Jiang, P., Chen, T., Luo, X., & Wen,
Q. 2017. A survey on the security of
blockchain systems. Future Generation
Computer Systems.
Lerer. M, 2019. The Taxation of
Cryptocurrency.Virtual Transactions Bring
Real-Life Tax Implications. Retrieved from:
https://commodity.com/cryptocurrency/bitco
in-cash/
Liu, W. 2018. Portfolio diversification
across cryptocurrencies. Finance Research
Letters (in press).
Lucey, B. M., Vigne, S. A., Ballester, L.,
Barbopoulos, L., Brzeszczynski, J.,
Carchano, O., ... Zaghini, A. 2018. Future
directions in international financial
integration research - A crowdsourced
perspective. International Review of
Financial Analysis, 55, 35–49.
Luther, W. J. 2016. Bitcoin Cash and the
future of digital payments. The Independent

FIN31
Review, 20(3), 397–404.
Luther, W. J., & Salter, A. W. 2017. Bitcoin
Cash and the bailout. The Quarterly Review
of Economics and Finance, 66, 50–56.
Maesa, D. D. F., Marino, A., & Ricci, L.
2017. Detecting artificial behaviours in the
Bitcoin Cash users graph. Online Social
Networks and Media, 3, 63–74.
MacNally, S. 2018. Predicting the Price of
Bitcoin Using Machine Learning. Retrieved
from:
https://www.researchgate.net/publication/32
5633087_Predicting_the_Price_of_Bitcoin_
Using_Machine_Learning
Megas. J. 2018. Opposing Bitcoin ABC and
Bitcoin SV Factions’ Debates Grow Heated
as the Bitcoin Cash Hard Fork Draws
Closer. Available
athttps://cointelegraph.com/news/opposing-
bitcoin-abc-and-bitcoin-sv-factions-debates-
grow-heated-as-the-bitcoin-cash-hard-fork-
draws-closer
Mira, N. 2018. An Analysis of
Cryptocurrency Governance
Retrieved from:
https://repository.upenn.edu/joseph_wharto
n_scholars/51/
Phyro. 2018.The history of ETC. Retrieved
from: https://medium.com/ethereum-
classic/the-history-of-etc-c6ded29e5831
Robert. W, 2018. Understanding the Budget
and Governance System. Retrieved from:
https://dashpay.atlassian.net/wiki/spaces/DO
C/pages/8585240/Understanding+the+Gove
rnance+and+Budget+Syste
Nikolaos A. K. 2019. A Survey on
Efficiency and Profitable Trading
Opportunities in Crypto currency Markets.
Journal of Risk and financial management ,
12, 67; doi:10.3390/jrfm12020067
Tasca, P., Hayes, A., & Liu, S. 2016. The
evolution of the Bitcoin Cash economy:
Extracting and analyzing the network of

FIN32
payment relationships. The Journal of Risk
Finance justaccepted 00-00.
Taylor S. &, Letham, B. 2017. Forecasting
at scale. Retrieved from:
https://peerj.com/preprints/3190/
Thies, S., & Molnar, P. 2018. Bayesian
change point analysis of Bitcoin Cash
returns. Finance Research Letters (in press).
Tiwari, A. K., Jana, R., Das, D., &Roubaud,
D. 2018. Informational efficiency of Bitcoin
Cash–an extension. Economics Letters, 163,
106—10 9.
Turk, ž., &Klinc, R. 2017. Potentials of
blockchain technology for construction
management. Procedia Engineering, 196,
638 —645.
Urquhart, A. 2016. The inefficiency of
Bitcoin Cash. Economics Letters, 148, 80–
82.
Urquhart, A. 2018. What causes the
attention of Bitcoin Cash? Economics
Letters, 166, 40–44.
Urquhart, A., & Zhang, H. 2018. Is Bitcoin
Cash a hedge or safe-haven for currencies?
An intraday analysis.
Valenzula. J ,2017. Trustlessness is
Effectively a Myth. Retrieved from:
https://www.dashforcenews.com/trustlessnes
s-effectively-myth/
Vandezande, N. 2017. Virtual currencies
under EU anti-money laundering law.
Computer Law & Security Review, 33(3),
341–353.
Vidal-Tomás, D., &Ibañez, A. 2018. Semi-
strong efficiency of Bitcoin Cash. Finance
Research Letters.
Vigne, S. A., Lucey, B. M., O’Connor, F. A.,
&Yarovaya, L. 2017. The financial
economics of white precious metals-A
survey. International Review of Financial
Analysis, 52, 292—30 8.
Vranken, H. 2017. Sustainability of Bitcoin
Cash and blockchains. Current Opinion in
Environmental Sustainability.
Zou, X. 2018.VECM Model Analysis of Carbon Emissions, GDP, and International

FIN33
Crude Oil Prices. Retrieved from:
https://www.hindawi.com/journals/ddns/201
8/5350308/