International Journal of Trend in Scientific Research and Development (IJTSRD)

Volume 7 Issue 2, March-April 2023 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470

Cryptocurrency Market Movement and Tendency Forecasting

using Twitter Emotion and Information Quantity
A. Esakki Elango1, E. Manohar ME, Ph.D2, S. Vishnu Durga3
1,3
ME Student, 2Professor,
1, 2, 3
Francis Xavier Engineering College, Tirunelveli, Tamil Nadu, India

ABSTRACT How to cite this paper: A. Esakki

Bitcoin was initially described to the public in a paper released in Elango | E. Manohar ME | S. Vishnu
2008 under the identity Satoshi Nakamoto. The first ever Bitcoin Durga "Cryptocurrency Market
transaction took place on January 3, 2009. Its success paved the way Movement and Tendency Forecasting
for the development of similar digital currencies in the years that using Twitter Emotion and Information
Quantity" Published
followed. There are more than 12,500 different cryptocurrencies,
in International
according to CoinMarketcap 2021. This is mostly owing to the Journal of Trend in
extraordinary volatility of the market, which drew many individuals Scientific Research
to take an interest and participate in it in the hopes of making money. and Development
Twitter has emerged as a common meeting place for those interested (ijtsrd), ISSN: 2456-
in cryptocurrencies. In a noteworthy move, Twitter announced on 6470, Volume-7 | IJTSRD55091
September 23, 2021, a new feature that would enable users to tip Issue-2, April 2023,
other users using their Bitcoin Lightning wallets. In spite of the fact pp.543-550, URL:
that this new technology may have far-reaching effects on our lives in www.ijtsrd.com/papers/ijtsrd55091.pdf
the future, there is not a great deal of writing on the subject of
cryptocurrencies. Even if there aren't many rules in place yet for Copyright © 2023 by author (s) and
International Journal of Trend in
trading cryptocurrencies, a social media sentiment study might help
Scientific Research and Development
fill in the gaps in our understanding of what influences bitcoin prices. Journal. This is an
In this study, we examine whether or not analyzing Twitter sentiment Open Access article
can reliably foretell changes in the value digital currencies. Seven of distributed under the
the most widely used cryptocurrencies have their own Twitter terms of the Creative Commons
discussions and price histories gathered. After that was done, the Attribution License (CC BY 4.0)
Valence Aware Dictionary for Sentiment Reasoning was used to (http://creativecommons.org/licenses/by/4.0)
conduct an analysis of the data's emotional content (VADER). We
used the Augmented Dicky Fuller (ADF), Kwiatkowski Phillips, KEYWORDS: Bitcoin, Augmented
Schmidt, and Shin (KPSS), and Granger Causality tests to identify DisckyFilter, Coin markup
time series that were stationary. However, the bullishness ratio
revealed that Ethereum and Polkadot prices were predicted despite
the fact that swings in Bitcoin, Cardano, XRP, and DOGE prices tend
to vary attitude. At last, we use Vector Autoregression (VAR) to look
at the predictability of price returns, and we discover that two of the
seven cryptocurrencies can have their prices predicted with a high
degree of accuracy. Exactness of price forecasts for Polkadot and
Ethereum, respectively, was 99.17% and 99.67%.

1. INTRODUCTION:
It's no secret that the popularity of cryptocurrencies institutional providers (Corbet et al., 2019; Nakamoto,
has exploded recently (Corbet et al., 2019; Kraaijeveld 2008). For examples, see (Corbet et al., 2019) and
& De Smedt, 2020; Rane & Dhage, 2019; Sattarov et (Nakamoto, 2008).
al., 2020). However, cryptocurrency investments have been
This is in part because their innovative blockchain- deemed risky because of the potential for it to be
based technology was developed with the aim of utilized in illegal activities like money laundering and
decentralizing finance and eliminating the mediation Ponzi schemes (Kethineni & Cao, 2020).
requirements of the current financial system, which
In addition, the extreme absence of regulations and the
are virtually exclusively delivered by banks and other
ever-changing nature of new technology combine to

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 543
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
provide a very unstable setting (Corbet et al., 2019; A clear definition of cryptocurrencies is crucial for
Kethineni & Cao, 2020; Kraaijeveld & De Smedt, grasping the phenomenon's rise and explaining it to
2020). Some examples of these studies may be found nonbelievers.
in (Corbet et al., 2019; Kethineni & Cao, 2020;
"An electronic method of payment based on
Kraaijeveld & De Smedt, 2020).
cryptographic proof, instead of trust, that allows any
A constant stream of information and social media two willing participants to directly engage with each
posts is produced in this frenetic environment. other without the requirement for trusted third party,"
Cryptocurrency enthusiasts and investors alike often Nakamoto said in the Bitcoin whitepaper. So far as I
turn to microblogging services like Twitter and Reddit can tell (Nakamoto, 2008, p.1)
for news and updates. To cite this article: Kraaijeveld,
Lansky (2018) offers an alternative definition of
R., and De Smedt,
cryptocurrencies as distributed digital currencies that
To this end, it has been determined that the sentiment satisfy the following six criteria:
of social media sites like Twitter and Reddit may be 1. As a result, (1) the system may reach agreement
used as a reliable indicator of future changes in the without relying on any one person or body to do
price of bitcoin. so.
Sattarov et al. (2020); Shen, Urquhart, and Wang 2. The units of bitcoin are tracked and recorded as
(2019); Smuts (2019); Wooley et al. (2019); Gao, belonging to certain individuals.
Huang, and Wang (2021); Kraaijeveld and De Smedt
3. The system determines the parameters for
(2020); Lamon, Nielsen, and Redondo (2016);
generating and owning new bitcoin units if this
Sattarov et al.
feature is enabled.
Finally, research evaluating the link between social
4. The ownership of cryptocurrency units can only
media sentiment and price fluctuations need to
be proven cryptographically.
account for the growing usage of automated systems
to post to social media. 5. It is acceptable to engage in the transfer of
ownership of cryptocurrency units via trading. A
Reference: (Rebholz, Wise, & Xiao, 2018) (Rebholz,
statement of the transaction cannot be produced
Wise, & Xiao, 2018)
without evidence of ownership of the units.
2. Literature Review 6. If more than one transaction using the same
This section explains what cryptocurrency are and
cryptography units is planned to take place at the
how they have evolved to this point, including
same time, the system will only execute one of
relevant historical context and the most recent
those transactions.
findings in the area. We also discuss the possible
utility of Twitter sentiment classification for Since Bitcoin's debut in 2009, the Cryptocurrency
forecasting market movements. Market has seen an explosion of investment, as seen
by the enormous growth in its market capitalization
Cyptocurrencies and number of transactions. One more aspect to
The landscape of Cryptocurrencies consider is that there will only ever be a maximum of
Over the last few weeks, cryptocurrencies have 21 million Bitcoins in circulation (issued by miners)
skyrocketed in popularity as governments and big until the year 2140. However, it may be divided into
corporations across the world have begun to accept smaller amounts called "Satoshis" (SATS), with 1
them. Examples of their widespread use include El Satoshi equaling 0.00000001 Bitcoin. The market
Salvador's 7 September 2021 adoption of Bitcoin as capitalization of all cryptocurrencies in 2021
legal cash and Twitter's 23 the ability to tip other users (according to CoinMarketCap)
using Bitcoin Lightning wallets will be introduced in
According to CoinMarketCap, the total value of the
September 2021..
cryptocurrencies rose from $5.5 billion at the
In 2008, a document defining and describing Bitcoin beginning of the year to $820,2 billion on Jan 7, 2018.
was released under the alias Satoshi Nakamoto. Due (2021). The total market value of all cryptocurrencies
to its Proof-of-Work user-consensus method, Bitcoin was predicted to surpass $2.3 trillion on November
is secure and free from threats like double spending 9th, 2021, an increase of more than 418%.
and hacking. And this peer-to-peer approach was
meant to do away with the time and money spent on
middlemen like banks. According to (Nakamoto,
2008)

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 544
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
their anonymity-promoting aspects make them
vulnerable to abuse in illegal contexts including
financing terrorism, tax fraud, such as Ponzi schemes
(Kethineni & Cao, 2020). To back this up, users don't
have to provide any private data to use the system.,
that they are not restricted in the number of accounts
that they may have, and that they have access to tools
that make their identification difficult (Dyntu &
Dykyi, 2019).
Coincheck, a prominent cryptocurrency exchange,
was hacked for $530 million in 2018, while Mt. Gox
was hacked for $400 million in 2014, demonstrating
the need for increased security.
Media announcements add to the unpredictability of
cryptocurrency values in the present market climate,
Figure 1: Value of Bitcoin in US Dollar Coins.
as studied by Corbet et al. (2019), who track how the
Binance Application Program Interface (API).
announcement of plans to construct KODAKCoin
The World Gold Council puts the value of gold at raised the company's stock price.This was similarly
about US$11.3 trillion, a figure that may be used to the case when a purported cooperation between
get a sense of the scope of the market. It's interesting Walmart and Litecoin was announced on September
to see how other cryptocurrencies have flourished as 13, 2021, but swiftly debunked as false. The
Bitcoin's competitors. There are now over 12,500 announcements were first shared on the official
distinct cryptocurrencies (CoinMarketCap, 2021). The Litecoin Twitter account and then published by a
Proof-of-Work consensus technique used by Bitcoin communication services business. Shortly after the
requires a lot of resources, however these alternatives announcement, the price of Litecoin jumped by more
aim to solve other issues or give more efficient than 30 percent, only to sink again once the hoax was
solutions. Vitalik Buterin, the creator of Ethereum, has exposed (New York Times, 2021). The entire thing
identified three key challenges that must be overcome lasted little more than twenty minutes, and Walmart
in the development of blockchain networks. The quickly published a statement denying the validity of
'Blockchain Trilemma' describes the interplay between the news.
decentralization, security, and scalability in the
blockchain ecosystem. There is still no clear answer to
the problem of how to accomplish all of these goals
without incurring unnecessary costs. The market
capitalization of all cryptocurrencies in 2021
(according to CoinMarketCap).

Despite the undeniable rise in popularity of

cryptocurrencies, some have voiced concerns that
their anonymity-promoting aspects make them
vulnerable to abuse in illegal contexts including
financing terrorism, tax fraud, such as Ponzi schemes
(Kethineni & Cao, 2020). To back this up, users don't
Figure 2: As of October 2021, these are the three have to provide any private data to use the system.,
cryptocurrencies with the most market that they are not restricted in the number of accounts
capitalization. CoinMarketCap is the data source. that they may have, and that they have access to tools
Despite the undeniable rise in popularity of that make their identification difficult (Dyntu &
cryptocurrencies, some have voiced concerns that Dykyi, 2019).

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 545
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
Coincheck, a prominent cryptocurrency exchange, Granger Causality testing was carried out to examine
was hacked for $530 million in 2018, while Mt. Gox the timing of events in a set of data. When all was said
was hacked for $400 million in 2014, demonstrating and done, VAR was used to make the final
the need for increased security. predictions.
Media announcements add to the unpredictability of
cryptocurrency values in the present market climate,
as studied by Corbet et al. (2019), who track how the
announcement of plans to construct KODAK Coin
raised the company's stock price. This was similarly
the case when a purported cooperation between
Walmart and Litecoin was announced on September
13, 2021, but swiftly debunked as false. The
announcements were first shared on the official
Litecoin Twitter account and then published by a
communication services business. Shortly after the
announcement, the price of Litecoin jumped by more
than 30 percent, only to sink again once the hoax was
exposed (New York Times, 2021). The entire thing
lasted little more than twenty minutes, and Walmart
quickly published a statement denying the validity of
the news.
3. Methodology
Research Model
Specifically, the study followed the stages in Figure 4
to conduct their investigation.
Using Twint for collecting tweets and the Binance
API for collecting cryptocurrency values, the first
datasets were acquired as part of the data gathering
process.
The data was then cleaned, and VADER was used to
analyze the opinions contained within. Next, we
combined tweets and price data and ran stationarity Figure 4: An Outline of the Research
tests using the ADF and the KPSS. Methodology
4. Collection of Data
Data Mining on Twitter
Initially, tweets on the most talked-about cryptocurrency were collected via programmatic interface for Twitter.
Twint is an open-source library that utilized because it is one of several Python libraries that can be used to gather
data from Twitter and because it can be used to circumvent restrictions imposed by the Twitter API. This allowed
for a large quantity of data to be gathered over a period of two months (01/09/21 - 01/11/21) pertaining to seven
of the most widely used cryptocurrencies. Several cryptocurrencies were chosen, including Bitcoin, Ethereum,
Cardano, Xrp, Solano, Doge, and Polkadot.
.csv files were created for each dataset. There was an option to change the language, and English was chosen.
Cryptosystem Tweets for Raw Data Data for Retweets
Bitcoin (BTC) 1.467.271 7.608.410
Ethereum (ETH) 575.773 1.484.263
Solana (SOL) 675.730 2.784.462
Cardano (ADA) 266.789 1.527.898
XRP (XRP) 268.284 864.578
Doge (DOGE) 353.573 1.085.648
Polkadot (DOT) 75.397 184.748
Total 3,682.817 15.540.007
Table 1: Total Tweets Collected

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 546
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
Date time open high low close Volume
19/11/2021 0:00:00 56891.62 57019.57 56876.99 57009.11 88.74734
19/11/2021 0:01:00 57005.07 57195.28 56999 57125.71 157.3485
19/11/2021 0:02:00 57120.01 57200 56885.16 56933.46 135.808
19/11/2021 0:03:00 56933.47 56942.8 56764.5 56803.89 52.84977
19/11/2021 0:04:00 56787.11 56829.43 56674 56763.4 47.95302
19/11/2021 0:05:00 56763.41 56850 56703.71 56743.24 58.1874
19/11/2021 0:06:00 56743.24 56867.83 56682.26 56839.7 48.90058
19/11/2021 0:07:00 56839.7 56954.53 56803.25 56921.21 26.23455
19/11/2021 0:08:00 56921.2 56931.47 56831.59 56915.53 21.98144
19/11/2021 0:09:00 56915.53 57150 56915.53 57143.95 43.94016
Table 2: Bitcoin Price Data sample
Id int64 To reach its conclusion, VADER uses a sentiment
conversation_id int64 dictionary, a collection of words that are often
created_at object classified as positive or negative based on their
Date object semantic orientation, as well as punctuation and
Time object emojis. VADER determines the positivity and
Timezone int64 negativity ratings, as well as the intensity with which
user_id int64 an emotion is positive or negative..
Username object A compound score is the total of individual lexical
Name object scores, scaled from -1 (very negative) to +1
Place object (highly positive) (extremely positive).
Tweet object If the sum of the scores is more than or equal to
0.05, then the mood is optimistic.
Table 3: Columns in Twitter Datasets
The middle ground is "Neutral Sentiment,"
5. Results and Discussions defined as a composite score of 0.05 or above that
VADER Sentiment Analysis includes both "compound score <-0.05" and
The tone of tweets is determined by using VADER "compound score -0.05."
sentiment analysis in this research. Pessimistic Opinions: (Overall Score = -0.05)
The "Valence Aware Dictionary for Sentiment In Table-8, you'll see VADER-scored tweets that have
Reasoning" is an example of a rule-based dictionary been cleaned using the aforementioned method. For
and sentiment analysis tool that focuses on social accurate scoring, tweets are tokenized and lemmatized
media sentiment. before they are sent.
Table 8
Tweet(tokenized/lemmatized) Compound negative neutral positive
['!', '!', 'big', 'fake', 'bubble', 'crash'] -0.7494 0.615 0.385 0
['disaster', 'waiting', 'happen', 'featured', 'news'] -0.6249 0.406 0.594 0

['following', 'closely', 'possible', 'bullish', 'opportunity', 'month', '.'] 0.4215 0 0.714 0.286
['future', 'wrapped', 'token', 'layer'] 0 0 1 0
['good', 'strong', 'project', '.'] 0.7351 0 0.326 0.674
['hard', 'ignore', 'even', 'though', 'bleeding', 'bit', ',', 'many', 'top',
0.6249 0.151 0.541 0.309
⎛ç ⎛', '◆ ⎛
[⎛
'strong', '.', 'good', 'sign', 'entire', 'market', '(', ')', '.']
['market', 'top', 'yet', 'ç
#
[
¡ [#
¡ ‘
• · ', 'ç ¡# [¡ç
#', "'", 'saying', "'", 'ç
ß̌ç̌
ß̌
0.7783 0 0.755 0.245
', '◆
:', '', '◆
- "
v
-v -
"◆']
[ 'true', 'believe', 'tap', 'get', 'extra', 'save', 'next', 'bullish', 'trend'] 0.7184 0 0.571 0.429
[':', 'dead', ',', 'bearish', 'month'] -0.6486 0.518 0.482 0
Table 8: VADER tweet scores

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 547
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
To compile the data for the sentiment analysis, we
used the '.describe()' function, the results of which are
displayed in Table 9. In the Appendix, you'll find
additional details about the final dataset used for each
cryptocurrency, daily graphs displaying their
cumulative score. Following figure represents the day-
day plot of bitcoin.
Using a composite score between zero and one
hundred, the tweets were sorted into the positive,
negative, and neutral categories. You can get a feel for
the daily volume of Bitcoin-related tweets and the
mood of the audience from Figure 13, which displays
the tweet volume and the tweet categories in which This study's datasets suggest that 28% of tweets are
they appear. Added to the Appendix are additional bar automated, hence this factor must be taken into
charts displaying the final rankings of each consideration when doing sentiment analysis.
cryptocurrency.
VADER, a sentiment analysis tool, proven to be
competent and capable of handling tweets, albeit
custom-made vocabulary, including regularly used
acronyms, had to be constructed to perform more
accurate sentiment analysis.
For the best outcomes, examine both the language
used and any cryptocurrency-specific acronyms.
The Granger Causality test findings demonstrate the
link between market mood and the cost of
cryptocurrencies. More study is needed to determine
the direction of causation between changes in price
and changes in emotion.
If a study reveals major discrepancies, it is important
to consider the unpredictability and incalculability of
Figure 13: Bitcoin Sentiment by Day market manipulation.
6. Discussion Despite the fact that this study's use of Vector
While this study's sample size was adequate, a higher Autoregression (VAR) only produced predictions that
number of tweets would have increased confidence in were quite close to the mark for two of the eight
the results and aided in the discovery of any obvious cryptocurrencies analyzed, further analysis of the time
regularity. The data set utilized in this study was series using different methods could reveal useful
decreased by 59.3 percentage points owing to the insights into which system is best suited to dealing
removal of irrelevant information and repetitions of with cryptocurrency data.
postings that were likely generated by automated
software. Future research might benefit from looking at other
cryptocurrencies using a broader data sample that
draws from a variety of social media platforms and
time periods.
Moreover, assessing user sentiment and pricing as a
means to detect probable market manipulation would
be an intriguing avenue to investigate.
7. Conclusion
Investors are drawn to the bitcoin market because it
represents a new and promising technology.
An effective analysis of user sentiment and research
into its predictive capacity is crucial in today's fast-
paced, data-driven society. Thus, the volatility of this
market, along with the widespread interest in

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 548
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
cryptocurrencies among social media users, presents a and Larisa Yarovaya. 2019. “Cryptocurrenciesas
rare opportunity. a Financial Asset: A Systematic Analysis.”
International Review of Financial Analysis
In this study, we contribute to the few prior research
62(September 2018):182–99.
on the topic by doing a literature review, sentiment
doi:10.1016/j.irfa.2018.09.003.
analysis, and an inquiry of the relationship between
social media data and price.. [7] Dickey, David A., and Wayne A. Fuller. 1979.
“Distribution of the Estimators for
The given methodology accounts for the fact that
Autoregressive Time Series With a Unit Root.”
Twitter bots make up roughly 28% of all tweets.
Journal of the American Statistical Association
These aided in the preprocessing of the tweets, along
74(366):427.doi:10.2307/2286348.
with the special vocabulary that was developed to deal
with abbreviations. [8] Dyntu, Valeriia, and Oleh Dykyi. 2019.
“Cryptocurrency in the System of Money
The short-term forecasts produced by this research are
Laundering.”
also quite accurate for two of the seven
cryptocurrency marketplaces analyzed. Using a Vector [9] Baltic Journal of Economic Studies 4(5):75.
Autoregression (VAR) model trained on Ethereum doi:10.30525/2256-0742/2018-4-5-75-81.
and Polkadot datasets spanning (2) months, we [10] Fama, Eugene F. 1970. “American Finance
produce an accurate forecast of 99.62% and 99.17%, Association Efficient Capital Markets : A
respectively, over the course of 12 hours. Reviewof Theory and Empirical Work Author (
The research already shows that social media s ): Eugene F . Fama Source : The Journal of
sentiment affects the bitcoin market, and our findings Finance , Vol . 25 , No . 2,Papers and
add to that. Proceedings of the Twenty- Eighth Annual
Meeting of the American.” The Journal of
If further developed, the approach used in this
Finance 25(2):383–417.
research might be utilized to generate prediction
models specific to marketplaces for cryptocurrencies [11] Gao, Xiang, Weige Huang, and Hua Wang.
more generally, or any market with similar features. 2021. “Financial Twitter Sentiment on Bitcoin
Return and High-Frequency Volatility.” Virtual
References
Economics 4(1):7–18.
[1] Antweiler, Werner, and Murray Z. Frank. 2004.
doi:10.34021/ve.2021.04.01(1).
“Is All That Talk Just Noise? The Information
Content of Internet Stock Message Boards.” [12] Giachanou, Anastasia, and Fabio Crestani.
Journal of Finance 59(3):1259–94. 2016. “Like It or Not: A Survey of Twitter
doi:10.1111/j.1540-6261.2004.00662.x. Sentiment Analysis Methods.” ACM Computing
Surveys 49(2). doi:10.1145/2938640.
[2] Baker, Malcolm, and Jeffrey Wurgler. 2007.
“Investor Sentiment in the Stock Market.” [13] Hao, Van Minh, Nguyen Huynh Huy, Bo Dao,
Journal of Economic Perspectives 21(2):129– Thanh Tan Mai, and Khuong Nguyen-An. 2019.
51. doi:10.1257/jep.21.2.129. [14] “Predicting Cryptocurrency Price Movements
[3] C. W. J. Granger. 1969. “Investigating Causal Based on Social Media.” Proceedings - 2019
Relations by Econometric Models and Cross- International Conference on Advanced
Spectral Methods.” Econometrica 37(3):424– Computing and Applications, ACOMP 201957–
38. 64. doi:10.1109/ACOMP.2019.00016.
[4] Colianni, Stuart, Stephanie Rosales, and [15] Ibrahim, Ahmed. 2021. “Forecasting the Early
Michael Signorotti. 2015. “Algorithmic Trading Market Movement in Bitcoin Using Twitter’s
of Cryptocurrency Based on Twitter Sentiment Sentiment Analysis: An Ensemble-Based
Analysis.” CS229 Project 1–5. Prediction Model.” 2021 IEEE International
[5] Corbet, Shaen, Charles Larkin, Brian Lucey, IOT, Electronics and Mechatronics
and Larisa Yarovaya. 2020. “KODAKCoin: A Conference, IEMTRONICS 2021 -
Blockchain Revolution or Exploiting a Potential Proceedings.doi:10.1109/IEMTRONICS52119.
Cryptocurrency Bubble?” Applied Economics 2021.9422647.
Letters 27(7):518–24. [16] Ibrahim, Ahmed, Rasha Kashef, Menglu Li,
doi:10.1080/13504851.2019.1637512. Esteban Valencia, and Eric Huang. 2020.
[6] Corbet, Shaen, Brian Lucey, Andrew Urquhart, “Bitcoin Network Mechanics: Forecasting the
BTC Closing Price Using Vector Auto-

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 549
 International Journal of Trend in Scientific Research and Development @ www.ijtsrd.com eISSN: 2456-6470
Regression Models Based on Endogenous and Cryptocurrency Markets.” Journal of Risk and
Exogenous Feature Variables.” Journal of Risk Financial Management 12(2):67.
and Financial Management 13(9):189. doi:10.3390/jrfm12020067.
doi:10.3390/jrfm13090189.
[23] Lamon, Connor, Eric Nielsen, and Eric
[17] Inamdar, Abid, Aarti Bhagtani, Suraj Bhatt, and Redondo. 2016. “Cryptocurrency Price
Pooja M. Shetty. 2019. “Predicting Prediction
Cryptocurrency Value Using Sentiment
[24] Using News and Social Media Sentiment.”
Analysis.” 2019 International Conference on
Pdfs.Semanticscholar.Org 25:96.
Intelligent Computing and Control Systems,
ICCS 2019 (Iciccs):932–34. [25] Lo, Andrew W. 2004. “The Adaptive Markets
doi:10.1109/ICCS45141.2019.9065838. Hypothesis.” Adaptive Markets 0920(617):176–
221. doi:10.2307/j.ctvc7778k.9.
[18] Kaplanski, Guy, and Haim Levy. 2010.
“Sentiment and Stock Prices: The Case of [26] Mohanty, Pavitra, Darshan Patel, Parth Patel,
Aviation Disasters.” Journal of Financial and Sudipta Roy. 2018. “Predicting
Economics 95(2):174–201. Fluctuations in Cryptocurrencies’ Price Using
doi:10.1016/j.jfineco.2009.10.002. Users’ Comments and Real-Time Prices.” 2018
7th International Conference on Reliability,
[19] Kethineni, Sesha, and Ying Cao. 2020. “The
Infocom Technologies and Optimization:
Rise in Popularity of Cryptocurrency and
Trends and Future Directions, ICRITO 2018
Associated Criminal Activity.” International
477–82. doi:10.1109/ICRITO.2018.8748792.
Criminal Justice Review 30(3):325–44.
doi:10.1177/1057567719827051. [27] Nakamoto, Satoshi. 2008. “Bitcoin: A Peer-to-
Peer Electronic Cash System.” SSRN Electronic
[20] Kraaijeveld, Olivier, and Johannes De Smedt.
Journal. doi:10.2139/ssrn.3440802.
2020. “The Predictive Power of Public Twitter
Sentiment for Forecasting Cryptocurrency [28] Narman, Husnu S., Alymbek Damir Uulu, and
Prices.” Journal of International Financial Jinwei Liu. 2018. “Profile Analysis for
Markets, Institutions and Money 65:101188. Cryptocurrency in Social Media.” 2018 IEEE
doi:10.1016/j.intfin.2020.101188. International Symposium on Signal Processing
and Information Technology, ISSPIT 2018
[21] Kwiatkowski, Denis, Peter C. B. Phillips, Peter
229–34. doi:10.1109/ISSPIT.2018.8642634.
Schmidt, and Yongcheol Shin. 1992. “Testing
the Null Hypothesis of Stationarity against the [29] Noda, Akihiko. 2019. “On the Time-Varying
Alternative of a Unit Root. How Sure Are We Efficiency of Cryptocurrency Markets.” Applied
That Economic Time Series Have a Unit Root?” Economics Letters 28(6):433–39.
Journal of Econometrics 54(1– 3):159–78. doi:10.1080/13504851.2020.1758617.
doi:10.1016/0304-4076(92)90104-Y. [30] Otoo, Maria Ward. 2005. “Consumer Sentiment
[22] Kyriazis. 2019. “A Survey on Efficiency and and the Stock Market.” SSRN Electronic
Profitable Trading Opportunities in Journal.

@ IJTSRD | Unique Paper ID – IJTSRD55091 | Volume – 7 | Issue – 2 | March-April 2023 Page 550