See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/382760407

Trading Bot for Cryptocurrency Market Based on Smart Price Action Strategies

Chapter · July 2024

DOI: 10.1002/9781394230952.ch8

CITATIONS READS

0 844

6 authors, including:

Divneet Singh Kapoor Kiran Jot Singh

Chandigarh University 38 PUBLICATIONS 497 CITATIONS
52 PUBLICATIONS 521 CITATIONS
SEE PROFILE
SEE PROFILE

Anshul Sharma
Chandigarh University
32 PUBLICATIONS 137 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kiran Jot Singh on 10 November 2024.

The user has requested enhancement of the downloaded file.

 8
Trading Bot for Cryptocurrency Market
Based on Smart Price Action Strategies
Divneet Singh Kapoor1,2*, Kiran Jot Singh1,2, Anshoom Jain3, Rhythm Chauhan3,
Khushal Thakur1,2 and Anshul Sharma1,2

Kalpana Chawla Centre for Research in Space Science & Technology, Chandigarh University,
1

Mohali, Punjab, India

2
Electronics and Communication Engineering Department, Chandigarh University,
Mohali, Punjab, India
3
Computer Science and Engineering Department, Chandigarh University, Mohali, Punjab, India

Abstract
This paper focuses on design of a Python-based crypto currency trading bot to support all major
exchanges and is managed through Telegram messenger application. The system offers a range of
capabilities, including back-testing, mapping, and money management, as well as machine learn-
ing strategy optimization. This enables users to create their own trading strategies and tactics using
Python and pandas. The strategies can be back-tested on historical data, which is widely and freely
available on exchange platforms and in the market that the user wants to trade in. The system also
allows for hyperoptimization of trading strategies based on the results of back-testing, which is
achieved using machine learning algorithms to find the optimal parameters for the strategy. In addi-
tion, the bot enables users to select markets to trade in by creating a static list or using an automated
one based on top traded volumes and/or prices. It also offers the ability to expressly ban marketplaces
in which the user does not want to trade. The system allows users to test their strategies with sim-
ulated money in dry-run mode or to deploy them with real money in live-trade mode. To monitor
trades, the bot can be integrated with the Telegram API. Users can display profit/loss, daily summa-
ries, current open trade outcomes, and many more options. Overall, the bot offers a powerful and
flexible platform for automating and optimizing crypto currency trading strategies.

Keywords: Cryptocurrency, frequency trade, cryptocurrency trading, back-testing, machine

learning

8.1 Introduction
Cryptocurrency trading has been a boon since 2020 and as good are its profits there are
losses of the same magnitude [1]. To minimize its losses one has to sit and analyse the
trends in market stay up to date in news, make predictions about a particular trade all while

*Corresponding author: [email protected]

Shubham Mahajan, Kapil Joshi, Amit Kant Pandit and Nitish Pathak (eds.) Integrating Metaheuristics in Computer Vision
for Real-World Optimization Problems, (151–162) © 2024 Scrivener Publishing LLC

151
152 Integrating Metaheuristics in Computer Vision

keeping the emotions aside which is a delicacy not many can accomplish. So to help with
this, this project is based on “automated crypto trading bots,” which will let even beginners
to trade and generate a passive income stream while minimizing losses. This bot will help
test the strategies made by you in historical data.
It is not physical currency like rupees or USD, but rather digital tokens based on block-
chain. Making it possible for people to conduct peer-to-peer transactions without the use of
banks [2]. That is, it is a decentralized kind of currency whose value is determined by what
individuals are willing to pay in the market for them. Bitcoin, Litecoin, Etherium, Tether,
Binance coin, Cardano, Doge, Solana, and WRX are examples of well-known cryptocurren-
cies [3]. In recent years, bitcoin market activity has surged significantly, and cryptocurrency
prices have climbed rapidly. The worldwide cryptocurrency market cap is expected to be
over $844 billion in 2022, with bitcoin having the greatest share of around $324 billion [4].
Figure 8.1 illustrates the crypto market cap acquired from April 28, 2013, to December 21,
2022, globally and as you can check it has been nearly $3,000,000,000,000 at its peak at the
end of 2021 and started capturing market during 2018 [5].
Figure 8.2 illustrates the share of each crypto coin in market of crypto-currency. Bitcoin
(BTC) (60% of the total market in 2022) has always captured the maximum market while
Solana (SOL) (~23% of the total market in 2022) of Anatoly Yakovenko captured the lowest
in the top ten coins [6].
The process of producing new digital “coins” for cryptocurrencies is known as mining.
It is, however, the extent of its simplicity. The procedure involves deciphering challenging
puzzles, validating bitcoin transactions on a blockchain network, and adding the verified
transactions to a distributed ledger in order to find these coins [7].
The process is not as easy as one may think.

• Cryptocurrency mining is the process of verifying new transactions and add-

ing them to the blockchain record on the cryptocurrency network.
• It is performed by solving complicated cryptographic hash problems to val-
idate transaction blocks that are updated on the decentralized blockchain
ledger.

Total Crypto Market Cap Chart

The chart below shows the total market cap & volume of cryptocurrencies globally, a result of 12,904 cryptocurrencies tracked across 618 exchanges.

24h 7d 14d 30d 60d 90d Max

Logarithmic Linear

Apr 29, 2013 Dec 21, 2022

$4,000,000,000,000

$3,000,000,000,000

$2,000,000,000,000

$1,000,000,000,000

$0.00

2014 2015 2016 2017 2018 2019 2020 2021 2022

2014 2016 2018 2020 2022

Figure 8.1 Total crypto market cap till 2022.

 Trading Bot for Cryptocurrency 153

Bitcoin (BTC) Dominance Chart

Chart below shows the bitcoin dominance percentage as compared to other cryptocurrencies in the top 10 ranking.

24h 7d 14d 30d 60d 90d Max

100%
Stacked Overlapping

80%
Dominance Percentage

60%

40%

20%

0%
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023

2014 2016 2018 2020 2022

BTC ETH USDT USDC BNB BUSD XRP DOGE ADA SOL Others

Figure 8.2 Individual cryptocurrency share in the market cap (2022).

To solve these issues, tremendous computational power and advanced equipment are
required. Miners are rewarded with cryptocurrency in return, which is then released into
circulation, thus the name cryptocurrency mining.
Freqtrade is a bitcoin trading robot system that is open source and free to use. It can be
controlled through Telegram and is designed to work with all major exchanges. It includes
features for back-testing, mapping, and money management, as well as machine learning
strategy optimization [8]. With Freqtrade, you can develop your own trading strategy in
Python using pandas, get historical market data for the exchange and markets you want to
trade in, back-test your strategy using past data, optimize your strategy by identifying the
best parameters with hyper-optimization and machine learning algorithms, select the mar-
kets you want to trade in, and run your strategy in either simulated or live trading mode.
You can also use the optional “edge” module to find the best historical trade expectations by
market and control and monitor your trades through Telegram or a REST API. Freqtrade
also includes tools for analyzing your back-testing data or trading history, including auto-
matic standard graphs and the ability to load data into interactive settings.
Data science (data science) is a field of study that combines subject-matter expertise,
programming skills, and a basic understanding of mathematics and statistics to extract
usable insights from data [9]. Crypto-bot uses data science by analyzing the previous can-
dles or the real time data and make the choice to consider ordering a trade or not. The bot
is programmed to analyze the data from previous years in order to determine whether or
not it is a good investment. This involves running the data through various scenarios and
analyzing the results. If the profits are not meeting the desired mark, the bot will rerun the
data with different strategies in order to see if there is a way to improve the outcome.
This method of prediction is considered safe because it is based on real data from the
past, rather than relying on speculation or guesswork. By using this approach, investors can
make informed decisions about whether to invest in a particular trade or not, rather than
taking unnecessary risks. Overall, this system allows for more accurate and reliable predic-
tions, which can help investors to maximize their profits and minimize their losses.
154 Integrating Metaheuristics in Computer Vision

Trading cryptocurrency can be a challenging and complex endeavor, requiring a signif-

icant amount of time, observation, and knowledge. For students and other professionals
with limited free time, it can be difficult to manage their schedules and put in the nec-
essary effort and attention required for successful trading. Additionally, the emotions of
greed and fear, which are naturally present in the human brain, can interfere with rational
thinking and decision-making, leading to unexpected losses. It is important to invest time
and effort in learning trend analysis and conducting research before making a trade, in
order to make informed and well-thought-out decisions. However, this can be a hectic and
time-­consuming task, especially for those who may not have the necessary understanding
or resources to invest in it. Overall, trading cryptocurrency is not for everyone, and those
who are interested in it should be prepared to put in the necessary time and effort, and be
mindful of the potential risks and challenges involved.
The main objectives/contribution of the presented work can be summarized as follows:

• To help people trade cryptocurrency with less stress and increase their prof-
itability. By automating the process of monitoring market conditions and
executing trades, these bots can help to alleviate some of the mental and
emotional burden of manual trading, and allow traders to focus on other
aspects of their business.
• To reduce the impact of human emotions on trading decisions. As men-
tioned earlier, emotions such as greed and fear can impair rational thinking
and lead to costly mistakes. By automating the trading process, bots can help
to eliminate the influence of emotions on decision-making, leading to more
consistent and objective trades.

8.2 Background
Since the introduction of crypto trading there have been many who know its potential in
the market and are known to its limitations. To address this, many worked on automated
trading using various models that worked shadowing its limitations, potential and how can
it be made better. The research papers were informative and allowed for further under-
standing of the project better.

1. Aisha Peng et al. in their “automated cryptocurrency trading bot implement-

ing DRL” [10] research paper presents an automated cryptocurrency trad-
ing system that utilizes machine learning. The system has shown promising
results and has the potential to outperform similar systems. However, there
are areas for improvement, such as further exploration of training parame-
ters and the possibility of enhancing input data features. Future work could
involve testing different model architectures and conducting correlation anal-
ysis to optimize performance. Additionally, training the model on the histor-
ical data of specific cryptocurrencies may improve accuracy and adaptability.
2. Eric Han in his “intra-exchange cryptocurrency arbitrage bot” [11]
research paper states cryptocurrency is a digital asset that uses encryption
to regulate its generation and verify financial transactions. There are many
 Trading Bot for Cryptocurrency 155

cryptocurrencies available, which can lead to price discrepancies due to mar-

ket inefficiencies. These inefficiencies happen when prices do not accurately
reflect the value of an asset and are more common in the cryptocurrency
market due to low regulation, high speculation, and investor behavior. Trades
called arbitrages, which involve buying and selling the same asset and result-
ing in more of that asset, can help address some of these pricing discrepancies
and make the market more efficient. This project attempted to use arbitrages
to make a profit on the binance cryptocurrency exchange and improve mar-
ket efficiency, although it did not achieve significant profits. It did, however,
generate a few hundred dollars over several months.
3. Dr. Ceren et al. in their Proposition of a trading bot for cryptocurrency mar-
ket [12] research paper shared that Blockchain technology has spawned new
business opportunities and seen rapid growth in market value in recent years.
As cryptocurrencies gain popularity, trading them becomes more important
but also carries high risk due to market volatility. This study aims to design a
free cryptocurrency trading bot that can be used by both inexperienced and
experienced traders, utilizing MACD and RSI indicators. The aim is to create
a user-friendly bot that capitalizes on all trading opportunities that present
themselves within a day.
4. Dhruv Piyus Parikh et al. in their Designing of Intelligent Auto-Bot for Financial
System [13] research paper share that Cryptocurrency values can fluctuate
significantly in a short amount of time, making it challenging for traders to
track price movements unless they are constantly monitoring them. Our sys-
tem employs IoT technology to assist with trading practices by offering real-
time updates from a crypto API through an autonomous crypto-stock alert
system. This system can send alerts for all schemes and stock or be filtered to
only alert for specific coins or assets. We use Bolt IoT and Ubuntu Terminal
to integrate with the Bolt cloud and WiFi module and utilize LED and buzz-
ers for alerts. This system helps investors to stay informed about price fluctu-
ations or irregularities in the market.
5. Anderson Simon et al. in their Crypto Proxima: An analysis of autonomous
Bitcoin trading [14] research paper explains that This project examined the
potential for profit from autonomous bitcoin trading using three models: an
LSTM neural network, a technical analysis model based on moving averages,
and a random model. The results indicated that the LSTM model had poten-
tial profitability at high frequencies, and the MA model had some ability to
predict the market. The random interval model showed simulations were
normally distributed around half the market gain or loss. These findings sug-
gest that it is possible to profit from autonomous trading, but further research
is needed to determine the optimal parameter selection.
6. Robert A. Caulk et al. in their FreqAI: generalizing adaptive modeling for
chaotic time-series market forecast [15] shared that FreqAI was developed to
compare adaptive timeseries forecasting techniques on chaotic cryptocur-
rency market data. It uses the open-source software Freqtrade as a founda-
tion, which supports data collection, storage, and live trading, as well as data
analysis and visualization. FreqAI has a user-friendly interface for integrating
156 Integrating Metaheuristics in Computer Vision

external machine learning libraries for adaptive time-series forecasting

and includes custom algorithms and methodologies to enhance computa-
tional and predictive performance. It can be used with all Python-accessible
machine learning libraries to generate new research on live and historical
data. Furthermore, advances in technologies like Internet of things, Wireless
sensor networks and computer vision can be used to develop newer multi
domain solutions as well.

From the literature cited above, it can be concluded that the automation of cryptocur-
rency trading has some future and can be bought in working conditions via various libraries
like freqtrade from FreqAI to test various methods that may be one’s mind and also test
them in dry run without losing any real money in the run. Thus, minimizing losses in the
actual trade. Furthermore, advances in technologies like Internet of things, Wireless sen-
sor networks and computer vision can be used to develop newer multi domain solutions
[16–22].

8.3 Proposed Framework

This work used FreqAI connected with Binance exchange via API to gather historical data
and use that data to stimulate the strategy developed, in dry run that whether it would have

Telegram App front-end API

Push Massages to Telegram bot Web application giving access to Enables live trading bot to
via the Telegram API to provide see the status of the across all the access latest status, orders and
realtime notifications on order, various tabled positions from this paper
trades, triggers and messages trading bot
Technology - Python Django
Technology - Telegram python front-end framework Technology - Django REST
API API framework

Telegram Web app Link to live

API front-end REST API trading bot

Jobs
Indicators Managers the different
Calculation of jobs required to be run
indicators and triggers regularly - market scans,
Calculate re validation of markets
on candle data Bot DB Jobs and
indicators throwing error, etc
scripts
and triggers
Technology - Back
trader Library Technology - CRON on
Linux and python scripts
Bot DB
Market data Central store for all the candle
data, indicators, orders, Algorithms
Fetch public candle positions, etc in tables Trading Different combinations
data via exchange APIs Market
algorithms of indicators to enter
data Technology - PostgreSQL and order and exit trades, set
Technology - CCXT API database management orders and set limit and
python API
Managed by the Django app stop prices

Technology - Python
Classes

Figure 8.3 Architecture overview for the paper trading bot [source: https://mattgosden.medium.com/
tradingbot-series-architecture-for- a-trading-bot-ac2352508c82].
 Trading Bot for Cryptocurrency 157

worked or not and provide improvement tips. The architecture of the trading bot is illus-
trated in Figure 8.3. The strategy used three indicators TEMA, RSI, BOLLINGER BANDS
set to desired limits. The results can be seen using telegram bot connected via freqAI using
bot key on the cellular device.
The algorithm for enabling the bot for cryptocurrency profit trading prediction is as
follows:

Step 1: Get the freqtrade image by downloading the dockercompose.yaml

file.
Step 2: In the yaml file, select the development image.
Step 3: Create a configuration file with parameters such as whitelist, backlist,
volume pair list, static pair list, maximum profit, maximum loss, maximum
amount in wallet, maximum number of transactions, and so on.
Step 4: Develop or establish a plan for the bot.
Step 5: Use indicators [Bollinger bands, RSI (relative strength index), TEMA
(triple exponential moving average), SMA (simple moving average)] to con-
figure and set the desired strategy in the bot.
Step 6: Connect to Binance with its API key.
Step 7: Import historical data from the Binance exchange service.
Step 8: Use this historical data to backtest your technique.
Step 9: The bot will display you how your strategy would have performed if
you had used it during that trade interval, and it will hyper optimize your
strategy.

And your bot is ready to use.

The indicators used for the setting the desired outcome for the algorithm are:

TEMA. The total average is calculated by using three different exponential moving averages
of three different candle ranges. We are employing this indication since it reacts quickly to
market momentum, which will help the bot reduce losses and increase profits [23].
Bollinger bands. It is used to represent the standard deviation from the exponential mov-
ing average, which effectively shows us the market’s likely movement. This indicator is com-
monly configured with a standard deviation of two and an EMA set to the last 20 candles
[24].
RSI. This indicator is commonly used to determine market momentum in conjunction with
other indicators. The market range for this indicator is thirty to seventy. If the RSI is less
than thirty, the coin has been oversold and could be a good entry point to long the market.
If the RSI is greater than seventy, the coin has been overbought, indicating that a bearish
trend is about to begin [25].
Figure 8.4 shows the Indicators that we used via visual representation on the chart (chart
showing highs and lows in prices of a trade in real time).

Heikin–Ashi candle. Heikin-Ashi being a candlestick pattern approach that tries to

decrease market noise by producing a chart that indicates trend direction better than tra-
ditional candlestick charts. The drawback of Heikin-Ashi is that risk may be impacted by
the loss of certain pricing data due to averaging. Long down candles with minimal upper
158 Integrating Metaheuristics in Computer Vision
Published on TradingView.com, Dec 22, 2022 12:24 UTC+5:30
Bitcoin / TetherUS, 1D, BINANCE O16824.68 H16860.77 L16793.50 C16808.70 –15.97 (–0.09%) USDT
BB (20, close, 2, 0) 17035.76 17696.82 16374.71
TEMA (9) 16707.99 24000.00

22000.00

20000.00

17696.82
17035.76
16808.70
17:05:05
16707.99
16374.71

14000.00
RSI (14, close, SMA, 14, 2) 46, 38 48, 69 ø ø

60.00
48.69
46.38
40.00

20.00
Jun Jul Aug Sep Oct Nov Dec 2023 Feb 20

TradingView

Figure 8.4 A graph showing RSI, Bollinger bands, TEMA.

shadow indicate significant selling pressure, and long up candles with little or no lower
shadows indicate strong purchasing pressure [26].

The strategy. We will use strategy to create a snowball effect, which means taking little prof-
its for one week and limiting stop losses to a minimum. For example, for 2% profit we will
risk 4% of our asset. Our strategy will primarily rely on a mix of TEMA, RSI, and Bollinger
Bands to forecast market movement in our favor. As a result, profit is booked. If the Heikin-
Ashi candle is above the TEMA, which signals the uptrend, and the chart has bounced off
the lower BB with an RSI greater than 30, this is a buy signal (green candle) If the Heikin-
Ashi candle is red and the momentum is downward, this indicates that the lower part of the
red candle is hitting TEMA and the candle is in the top part of the Bollinger band, which is
fifty to seventy, this is a sell signal.

8.4 Results
The bot ran well in the DRY RUN (as shown in Figure 8.5) using the mentioned strategy,
without the help of any human interaction checking all the mentioned objectives. It was
easy to trade and test our strategy on the data from binance which allowed us to view the
scope of the improvements. The bot traded on its own and suffered losses of 51% having
balance of 9776.376 USDT in the end and showed us that this strategy may fail in the cur-
rent market based on previous trends, as depicted in Figure 8.6.
It can be seen in Table 8.1, that user suffered losses in our trade except the first one which
means we would have lost money if we used our strategy during that time in these trades so
we need to improve our strategy to make profit.
Table 8.2 summarizes the trades done on the historic data from binance, the bot made
three maximum trades at a time, and did 162 average trades per day and started with 20000
fake money on dry run and having final balance of 9776.376 USDT and suffering losses of
51.12%.
 Trading Bot for Cryptocurrency 159

Figure 8.5 Results from back test in DRY run.

SOL/USDT

250 rsi
Volume
Exit - Loss
200 Exit - Profit
Trade entry
sar
tema
150 Bollinger Band
Price

exit_long
enter_long
100 Price

50

0
Volume

1M
0.5M
0

80
60
RSI

40
20
Jan 2022 Mar 2022 May 2022 Jul 2022 Sep 2022 Nov 2022 Jan 2023

Figure 8.6 The result of this paper’s applied strategy on the historical data from Binance.
160 Integrating Metaheuristics in Computer Vision

Table 8.1 THE BACKTEST REPORT, applied our strategy used on historical data from binance.
Backtesting report

Avg Cum Tot

profit profit Tot profit profit Avg
Pair Entries % % USDT % duration Win Draw Loss Win%

SOL/USDT 18 0.29 5.14 514.210 2.57 6:43:00 8 4 6 44.4

MATIC/USDT 36 −0.50 −18.08 −1809.421 −9.05 7:59:00 21 6 9 58.3

BTC/USDT 46 −0.47 −21.55 −2156.897 −10.78 8:44:00 18 6 22 39.1

ATOM/USDT 29 −0.90 −25.97 −2599.332 −13.00 8:34:00 10 1 18 34.5

GALA/USDT 33 −1.26 −41.68 −4172.184 −20.86 10:26:00 11 5 17 33.3

TOTAL 162 −0.63 −102.14 −10223.624 −51.12 8:40:00 68 22 72 42.0

Table 8.2 This table shows the whole summary of our trade.
Metric Value
Backtesting from 2021−12−02 15:00:00
Backtesting to 2022−12−22 04:00:00
Max open trades 3

Total/Daily Avg Trades 162 / 0.42

Starting balance 20000 USDT
Final balance 9776.376 USDT
Absolute profit −10223.624 USDT
Total profit % −51.12%
CAGR % −49.36%
Profit factor 0.56
Trades per day 0.42
Avg. daily profit % −0.13%
Avg. stake amount 9999.761 USDT
Total trade volume 1619961.362 USDT

Best Pair SOL/USDT 5.14%

Worst Pair GALA/USDT −41.68%
Best trade GALA/USDT 8.07%
Worst trade MATIC/USDT −30.34%
Best day 959.978 USDT
Worst day −2556.608 USDT
Days win/draw/lose 35 / 5 / 45
Avg. Duration Winners 5:14:00
 Trading Bot for Cryptocurrency 161

8.5 Conclusion and Future Scope

The proposed work provided a brief overview of crypto currency, crypto mining and how to
automate a hectic task of crypto trading with the help of freqtrade a python-based library.
The work provides the whole framework of the project from the libraries used to the indi-
cators used during the formation of the trading strategy. After analysing the data one can
make changes to the strategy. Thus it is advised to test the strategy and working of the bot
in dry run before using actual money trade. Automated crypto trading via programmed
bots such as Freqtrade has the potential to significantly alter the future of cryptocurrency
trading. These bots are intended to automate the purchasing and selling of bitcoin based
on predefined rules and strategies. Traders can save time and effort while potentially mak-
ing more educated and lucrative transactions by deploying automated trading bots. Using
automated crypto trading bots has various potential advantages as it can help to decrease
the impact of emotions on trading decisions since they adhere to a set of rules rather than
being persuaded by emotions such as greed or fear. Furthermore, automated trading bots
may examine market data and patterns in real time, potentially uncovering trading oppor-
tunities that a human trader would overlook. This is especially important for traders who
do not have the time or resources to monitor the market on a continuous basis.
It is crucial to remember, however, that automated crypto trading bots are not without
hazards and limitations. They are susceptible to errors or malfunctions, and they may not
always make the most profitable deals. Furthermore, they may be unable to respond to
unforeseen market occurrences or changes in market conditions. As a result, in order to
maximize their chances of success, traders must carefully assess the potential benefits and
hazards of utilizing automated trading bots, and use them in conjunction with other meth-
ods and tools.

References
1. Yu, Z., Razzaq, A., Rehman, A., Shah, A., Jameel, K., Mor, R.S., Disruption in global supply chain
and socio-economic shocks: A lesson from COVID-19 for sustainable production and con-
sumption. Oper. Manage. Res., 15, 1–2, 233–248, Jun. 2022, doi: 10.1007/S12063-021-00179-Y/
FIGURES/9.
2. Taskinsoy, J., Say good bye to physical cash and welcome to central bank digital currency. SSRN
Electron. J., Nov. 2021, doi: 10.2139/ssrn.3972858.
3. Tredinnick, L., Cryptocurrencies and the blockchain. Bus. Inf. Rev., 36, 1, 39–44, Mar. 2019, doi:
10.1177/0266382119836314.
4. Robertson, K. and Zhang, J., Suitable price discovery measurement of bitcoin spot and futures
markets. SSRN Electron. J., Jan. 2022, doi: 10.2139/ssrn.4012165.
5. Biju, A.V., Merin Mathew, A., Nithi Krishna, P.P., Akhil, M.P., Akhil akhil, M.P., Is the future
of bitcoin safe? A triangulation approach in the reality of BTC market through a sentiments
analysis. Digit. Finance, 44, 4, 275–290, Jun. 2022, doi: 10.1007/S42521-022-00052-Y.
6. Bhujel, S. and Rahulamathavan, Y., A survey: Security, transparency, and scalability issues of
NFT’s and its marketplaces. Sensors, 22, 22, 8833, Nov. 2022, doi: 10.3390/s22228833.
7. Tosh, D.K., Shetty, S., Liang, X., Kamhoua, C., Njilla, L., Consensus protocols for blockchain-­
based data provenance: Challenges and opportunities, in: 2017 IEEE 8th Annual Ubiquitous
 162 Integrating Metaheuristics in Computer Vision

Computing, Electronics and Mobile Communication Conference (UEMCON), vol. 2018-Janua,

pp. 469–474, Oct. 2017, doi: 10.1109/UEMCON.2017.8249088.
8. Loh, L.K.Y., Kueh, H.K., Parikh, N.J., Chan, H., Ho, N.J.H., Chua, M.C.H., An ensembling
architecture incorporating machine learning models and genetic algorithm optimization for
forex trading. FinTech, 1, 2, 100–124, Mar. 2022, doi: 10.3390/fintech1020008.
9. Boehmke, B., Hazen, B., Boone, C.A., Robinson, J.L., A data science and open source software
approach to analytics for strategic sourcing. Int. J. Inf. Manage., 54, 102167, Oct. 2020, doi:
10.1016/j.ijinfomgt.2020.102167.
10. Peng, A., Ang, S.L., Lim, C.Y., Automated cryptocurrency trading bot implementing DRL.
Pertanika J. Sci. Technol., 30, 4, 2683–2705, Sep. 2022, doi: 10.47836/pjst.30.4.22.
11. Han, E., Intra-exchange cryptocurrency arbitrage bot, San Jose State University, San Jose, CA,
USA, 2018.
12. Cubukcu Cerasi, C., Kir, T., Ince, D., Proposition of a trading bot for cryptocurrency market.
Soc. Sci. Dev. J., 7, 33, 237–243, Sep. 2022, doi: 10.31567/ssd.726.
13. Parikh, D.P., Singh, R., Banerjee, A., Design of intelligent auto-bot for financial system. Int. J.
Adv. Res. Sci. Commun. Technol., 3, 2, 283–289, Mar. 2021, doi: 10.48175/IJARSCT-910.
14. Andersson, S., Häggström, J., Icimpaye, U., Lindstedt, H., Crypto proxima: An analysis of
autonomous bitcoin trading, 2021.
15. Caulk, R.A. et al., FreqAI: Generalizing adaptive modeling for chaotic time-series market fore-
casts. J. Open Source Softw., 7, 80, 4864, Dec. 2022, doi: 10.21105/joss.04864.
16. Singh, K.J. et al., Adaptive flower pollination algorithm-based energy efficient routing protocol for
multi-robot systems. IEEE Access, 9, 82417–82434, 2021, doi: 10.1109/ACCESS.2021.3086628.
17. Sachdeva, P. and Singh, K.J., Automatic segmentation and area calculation of optic disc in oph-
thalmic images, in: 2015 2nd International Conference on Recent Advances in Engineering &
Computational Sciences (RAECS), pp. 1–5, Dec. 2015, doi: 10.1109/RAECS.2015.7453356.
18. Singh, K.J., Kapoor, D.S., Thakur, K., Sharma, A., Gao, X.Z., Computer-vision based object
detection and recognition for service robot in indoor environment. Comput. Mater. Contin., 72,
1, 197, Feb. 2022, doi: 10.32604/CMC.2022.022989.
19. Sharma, A., Kapoor, D.S., Nayyar, A., Qureshi, B., Singh, K.J., Thakur, K., Exploration of IoT
nodes communication using LoRaWAN in forest environment. Comput. Mater. Contin., 71, 2,
6240–6256, 2022, doi: 10.32604/CMC.2022.024639.
20. Jawhar, Q., Thakur, K., Singh, K.J., Recent advances in handling big data for wireless sensor
networks. IEEE Potentials, 39, 6, 22–27, 2020.
21. Sandhu, H.S., Singh, K.J., Kapoor, D.S., Automatic edge detection algorithm and area calcula-
tion for flame and fire images. Proc. 2016 6th Int. Conf. - Cloud Syst. Big Data Eng. Conflu. 2016,
pp. 403–407, Jul. 2016, doi: 10.1109/CONFLUENCE.2016.7508152.
22. Singh, K.J., Kapoor, D.S., Abouhawwash, M., Al-Amri, J.F., Mahajan, S., Pandit, A.K., Behavior
of delivery robot in human-robot collaborative spaces during navigation. Intell. Autom. Soft
Comput., 35, 1, 795–810, 2023.
23. Marshall, B.R., Young, M.R., Rose, L.C., Candlestick technical trading strategies: Can they
create value for investors? J. Bank Financ., 30, 8, 2303–2323, Aug. 2006, doi: 10.1016/j.
jbankfin.2005.08.001.
24. Macedo, L.L., Godinho, P., Alves, M.J., Mean-semivariance portfolio optimization with mul-
tiobjective evolutionary algorithms and technical analysis rules. Expert Syst. Appl., 79, 33–43,
Aug. 2017, doi: 10.1016/j.eswa.2017.02.033.
25. Hassen, O.A., Darwish, S.M., Abu, N.A., Abidin, Z.Z., Application of cloud model in qualitative
forecasting for stock market trends. Entropy, 22, 9, 991, Sep. 2020, doi: 10.3390/e22090991.
26. Di Lorenzo, R., Heikin Ashi, pp. 165–169, Springer, Milano, 2012.

View publication stats