International Journal of Scientific Research in Science and Technology
Available online at : www.ijsrst.com
Print ISSN: 2395-6011 | Online ISSN: 2395-602X doi : https://doi.org/10.32628/IJSRST
Machine Learning-Based Analysis of Crypto Currency Market
Financial Risk Management
P Deepthi, 2Jella Shreya, 3Vangala Snehitha
1
1 Associate Professor, Department of CSE, Bhoj Reddy Engineering College for Women, Hyderabad, Telangana,
India
2,3Students, Department of CSE, Bhoj Reddy Engineering College for Women, Hyderabad, Telangana, India
ARTICLEINFO ABSTRACT
Crypto currency is a form of digital currency that relies on cryptography to
Article History:
maintain and verify transactions, instead of a centralized authority.
Accepted: 01 July 2023
However, this decentralized nature can lead to several risks that can
Published: 12 July 2023
impact the assessments carried out by risk auditors. Money laundering is a
significant financial risk associated with the growing popularity of crypto
currency. This paper proposes machine learning - based approach that uses
Publication Issue
Hierarchical Risk Parity and unsupervised machine learning to analyse the
Volume 10, Issue 4
financial risk associated with crypto currency markets. The study finds
July-August-2023
that machine learning algorithms can effectively capture the complex
relationships between variables and provide accurate risk management.
Page Number
The study underscores the potential of machine learning based analysis to
106-110
improve financial risk management in the constantly evolving world of
crypto currencies.
Keywords: Crypto currency, Risk Management, Financial Risk, Money
Laundering
I. INTRODUCTION correlation matrix within a hierarchical structure, and
the highly volatile nature of crypto currency. The
The financial market is a complex system that has not value of crypto currency can fluctuate rapidly and
been widely accepted by universities in terms of its unpredictably, impacting both regulated and
definition of complexity, leading to disagreements on unregulated environments [7] -[10]. News outlets also
how to define and understand the interactions closely monitor price changes and significant market
between its elements. Modelling complex systems is a movements. Rules and regulations have been
daunting task, as they are structured hierarchically implemented to safeguard investors and prevent
with their own subsystems, and these hierarchical money laundering [11]. To manage tail risk and
models are used to extract resources. Portfolio achieve positive outcomes, [11] proposes the use of
construction can be challenging due to the lack of a the Hierarchical Risk Parity (HRP) strategy for multi
Copyright © 2023 The Author(s): This is an open-access article distributed under the terms of the Creative 106
Commons Attribution 4.0 International License (CC BY-NC 4.0) which permits unrestricted use, distribution,
and reproduction in any medium for non-commercial use provided the original author and source are credited.
P Deepthi et al Int J Sci Res Sci & Technol. July-August-2023, 10 (4) : 106-110
- asset, multi – factor allocation. Similarly, Jain et al. statements. Identifying the crypto currency’s highest
[12] utilized the HRP strategy to analyse fifty indexes probability risk.
of NIFTY stocks.
According to the 2020 report by Chartered
Professional Accountants Canada (CPA Canada),
investing in Crypto currencies poses several unique
risks, including:
Crypto currencies are not subject to the same
regulations as traditional investments, which can
make them more susceptible to fraud, market
manipulation, and other illegal activities. Figure 1 : Comparison of the recent cut - edge in
Investors should be aware that crypto currencies are Cryptography risk analysis.
known for their extreme volatility, making them
susceptible to sudden and unpredictable price changes II. RELATED WORK
that could result in substantial financial losses.
Crypto currencies are stored electronically, and Money laundering using crypto currencies involves
therefore, are vulnerable to hacking, theft, and other using the technology to conceal the origin of illegal
cyber - attacks. Investors must take appropriate funds. Crypto currencies have been attractive to
measures to protect their crypto currency holdings. money launderers because they can provide a level of
Despite the growing popularity of crypto currencies, anonymity and can be traded easily across borders.
they are not widely accepted as a form of payment, Financial institutions face challenges in combating
which can limit their usefulness and adoption. crypto currency - related crimes due to the advanced
Unlike traditional investments, such as stocks and technological features of crypto currencies, which can
bonds, Crypto currencies do not have an intrinsic be difficult to understand for banks and regulatory
value, which makes it difficult to determine their true agencies. This knowledge gap can create barriers in
worth. Crypto currencies have made a significant the fight against money laundering and other illicit
impact in both regulated and unregulated activities carried out using crypto currencies. Figure 2
environments, with news outlets closely monitoring presents study of money laundering cases in India and
price fluctuations and market movements. Regulatory Figure 3 presents a comparison of the market
frameworks have been established to protect investors capitalization of six different crypto currencies as of
and prevent money laundering. To manage risk and March 2023 as per Coin Market Cap.
produce favorable outcomes, the
Hierarchical Risk Parity (HRP) strategy is
recommended for allocating multiple assets and
factors. In summary, the main contribution of this
study is:
Implementation of the Hierarchical Risk Parity
approach for crypto currency portfolio management
using machine learning techniques. The proposed
system analyses the professional accounting
perspective by assessing the associated risks of crypto
currency and the anticipated impact on financial
Figure 2 : Case Study of Money Laundering in India
International Journal of Scientific Research in Science and Technology (www.ijsrst.com) | Volume 10 | Issue 4 107
P Deepthi et al Int J Sci Res Sci & Technol. July-August-2023, 10 (4) : 106-110
It is true that financial institutions and banks are
often primary targets for money laundering activities.
When banks are involved in money laundering, it can
damage their reputation and erode the trust of Through the recursive approach, clusters are formed
customers, who may become concerned about the based on Equation 2. The initial cluster is designated
safety of their funds. When a bank experiences a loss as the first element in the set of clusters, denoted as C.
of confidence, it may result in a decrease in its overall
value. This situation may arise due to various reasons
such as financial instability, poor management, or Using this method, the distance matrix is updated for
fraudulent activities. Similarly, customers may the (2) evaluation process, and all assets use the single
become more hesitant to use the bank's services if clustering linkage C [1]. As a result, for any asset x
they perceive that it is not taking adequate measures that is not part of the cluster, the distance of the new
to prevent money laundering. Therefore, it is cluster is calculated according to Equation 4:
important for financial institutions and banks to have
robust anti – money laundering policies and
procedures in place to prevent and detect money
laundering activities.
Figure 4: Data
The data used consists of daily crypto currency prices
Figure 3: Market Capability of Crypto currency as of from 2017 to 2020, obtained from the coin market cap.
March 2023 To ensure compatibility with the applied algorithm,
III. PROPOSED SYSTEM missing data information was excluded and reliable
This section provides details of the proposed approach forwarded observations were used to fill in any gaps.
for predicting exchange rates, which utilizes the The total dataset contains ten thousand records, with
graph – based theory of HRP and machine learning 80% allocated for the training set and 20% for the
techniques. The approach involves three key steps: testing set. Figures 3, 4, 5, and 6 provide evidence of
clustering, recursive bisection, and quasi - significant growth during 2016 and 2017, followed by
diagonalization. The first step includes clustering a sharp decline.
assets using the Hierarchical Tree Clustering
algorithm. The correlation matrix between two assets IV. RESULTS AND ANALYSIS
x and y is converted to the correlation distance matrix
A using Equation 1.: Out of all traditional well - known approaches like
risk - based asset allocation strategies: Inverse
The subsequent stage involves evaluating the pair Volatility (IV), Minimum Variance (MV), and
wise Euclidean distance between columns, which Maximum Diversification (MD), using HPR gave the
results in the generation of the augmentation matrix optimal results. HPR portfolio used a 350 - day
distance, as demonstrated by Equation 2. covariance estimation. The HPR annualized volatility
International Journal of Scientific Research in Science and Technology (www.ijsrst.com) | Volume 10 | Issue 4 108
P Deepthi et al Int J Sci Res Sci & Technol. July-August-2023, 10 (4) : 106-110
and return were 0.7718 and 1.7802, respectively. This study aims to analyze the risk management of
Comparing the results with the other traditional Crypto currency networks by utilizing the
approaches, the balance between risk and return of Reinforcement Learning (RL) technique and the
HPR had a significant impact, as it provided the best Hierarchical Risk Parity (HRP) asset allocation
trade - off between risk and return. method in a crypto currency portfolio. RL produced
high - performance evaluation results compared to
other machine learning techniques used in this field.
The learning - based approach of RL makes it suitable
for providing accurate information in this process.
HRP was chosen due to its desirable diversification
and properties. The results were analyzed using
various estimation windows and methodologies, and
the selected period was rebalanced similarly. The
study proposes future research to extend this
technique by conducting out - of - sample testing
performance on more assets and classes and using
optimization techniques to identify risk evaluation
Figure 5: Reinforcement Learning based architecture
and improve risk management performance. The
for risk management
keywords for this study are crypto currency, risk
management, Reinforcement Learning, Hierarchical
Risk Parity, and asset allocation.
VI. REFERENCES
[1]. Lee, S. hyun. &Kim Mi Na, (2008) “This is my
paper”, ABC Transactions on ECE, Vol.10, No.5,
pp120 -122.
[2]. Gizem, Aksahya&Ayese, Ozcan (2009)
Coomunications& Networks, Network Books,
ABCPublishers.
[3]. Zheng, Y., &Liu, X. (2020). A Comprehensive
Review on Cryptocurrency Trading Risk
Management, Security and Investment. IEEE
Access, 8, 21512 -21523.
[4]. gob, P. L., &Tan, Y. (2020). A Study of
Cryptocurrency Risk Management Frameworks.
IEEE Access, 8, 69937 - 69947. Ryu, B. (2019).
[5]. Cryptocurrency and Cybersecurity pitfalls and
Figure 6: Crypto currency working results. Journal of Cybersecurity and
Information Management, 2 (1), 1 - 16.
V. CONCLUSION [6]. Gill, S. S., &Goyal, N. K. (2020). Cryptocurrency
Security pitfalls and Countermeasures.
International Journal of Scientific Research in Science and Technology (www.ijsrst.com) | Volume 10 | Issue 4 109
P Deepthi et al Int J Sci Res Sci & Technol. July-August-2023, 10 (4) : 106-110
International Journal of Computer Applications, Cite this article as :
179 (25), 10 - 14.
[7]. Huang, D. (2019). Cryptocurrency Investment P Deepthi, Jella Shreya, Vangala Snehitha, "Machine
and Risk Management A Survey. Journal of Risk Learning-Based Analysis of Crypto Currency Market
and Financial Management, 12 (6), 302 - 317. Financial Risk Management", International Journal of
[8]. Li, K., &Li, K. (2020). Cryptocurrency Investment Scientific Research in Science and Technology
finances A New Asset Class. Journal of (IJSRST), Online ISSN : 2395-602X, Print ISSN : 2395-
Alternative Investments, 23 (3), 51 - 63. 6011, Volume 10 Issue 4, pp. 106-110, July-August
[9]. Dempster, M. A., Liu, C., &Tang, A. (2021). 2023.
Pricing cryptocurrency collective finances using Journal URL : https://ijsrst.com/IJSRST5231048
sentiment analysis. Journal of Financial Data
Science, 3 (3), 97 -109.
[10]. Roubaud, R., Nguyen, D. K., &Reboredo, J. C.
(2021). threat and return nexus in
cryptocurrency requests substantiation from
literal data. Journal of Risk and Financial
Management, 14 (5), 226 - 242.
[11]. Lucey, M. K., Zhang, C., Dowling, S., &Urquhart,
A. (2019). Cryptocurrency request comovements
A unproductive analysis.
[12]. Bose, T., &Saha, S. (2021). Price discovery in
Bitcoin spot and futures requests A cointegration
analysis. Economics Letters, 203, 109733. Hatemi
- J, S. M., Irandoust, F., &Shaukat, A. M. (2021).
[13]. The impact of COVID - 19 on Bitcoin volatility
and safe - haven parcels substantiation from the
GARCH model. Finance Research Letters, 42,
101757.
[14]. Yu, C. L. (2020). Bitcoin and energy consumption
A quantitative analysis. Energy Policy, 146,
111889. Bouoiyour, H., Selmi, M., &Tiwari, F.
(2019).
[15]. Is Bitcoin bullish or bearish? A new perspective
through energy consumption. Energy Economics,
84, 104660.
International Journal of Scientific Research in Science and Technology (www.ijsrst.com) | Volume 10 | Issue 4 110
