Introduction
The technology of artificial intelligence (A.I.) has the potential of transforming many aspects of humanity’s life, including the existing financial operations. The emergence of digital currencies, for instance, bitcoin, in 2008 presented a question of analyzing this new financial market. This paper aims to review the application of A.I. in the context of blockchain finance by examining scholarly articles to determine whether the A.I. algorithm can be used to analyze this financial market.
Overview
The widespread use of cryptocurrencies and the fact that they remained popular for over ten years facilitates the need for developing prediction models that will allow one to use these currencies as investments. A.I. technology has been developing for decades and has been a topic of discussion for many. It has implications for presenting an efficient analysis and presenting one with a practical evaluation of this financial market. However, the complexity of Blockchain and the nature of this technology makes it impossible to analyze and predict using the standard financial analysis models. Therefore, behavioral factors and other elements should be accounted for in the successful algorithm.
Research Objectives
This research aims to examine peer-reviewed articles and determine the applicability of the A.I. algorithm and data mining and its implications for the cryptocurrency market. The amount of data generated by exchange services and activities allow researchers to analyze behavior and trends and come up with valid prediction models. An essential factor that will be examined in this paper is emotions as a critical element in the decision-making process of individuals. In this regard, the correlation between social media and news outlets and its impact on the fluctuations of cryptocurrencies will be examined as well.
Thus, the main focus of this paper is blockchain technology, A.I., and cryptocurrencies. Additionally, the ability of A.I. to assess and evaluate this financial market in order to boost is one of the objectives. Understanding different models of A.I. and their results in regard to cryptocurrency market predictions will be explored. Next, a test with the use of historical data will be conducted, and results, as well as their implications will be presented. In general, this paper should help develop an ethical framework for decision-making in the context of A.I. facilitated cryptocurrency market analysis.
Research Questions
The fact that A.I. algorithms are free from bias that data analytics can suffer from presents an implication for applying this technology to make accurate forecasting for the cryptocurrency market. This research aims to answer the question of plausibility and validity that A.I. has in regards to financial market analysis. Additionally, the ethical implications of this technology will be examined as part of the study.
Research Methodology
In order to locate answers for the questions that this paper aims to answer, peer-reviewed articles from scholarly journals that focus on A.I., finance, cryptocurrency, and Blockchain were examined. The reports present a large number of data, including background information and specific studies that use A.I. to make financial predictions. Thus, this paper will offer an assessment of the definition of Blockchain, bitcoin, and A.I. Additionally, the connection between these elements will be explored. The limitations of Blockchain will be investigated as well as other aspects of this technology. A variety of models and strategies of A.I. application will be reviewed in order to determine a successful strategy for A.I. implementation. This information should provide a cohesive understanding of the cryptocurrency market, and the implications of A.I. enabled forecasting.
Literature Review
The following paragraphs will focus on exploring the research works that examine various aspects of Blockchain and A.I. It should be noted that the development of a framework that can be used to analyze the financial market of cryptocurrencies is critical due to the significant impact of blockchain technology on the economic and social life of people globally. Machine learning algorithms that will be discussed in this paper are an essential element of A.I. technology, and some examples include neural networks and deep learning. Due to the current popularity of cryptocurrencies and their impact on the economy on a global level, a need for forecasting the trends and specifics of value changes has arisen.
Background and Related Works
Firstly, it is crucial to define A.I. as well as blockchain technology and describe the application of the former in the financial markets. According to Zheng et al. (2018, p. 1), A.I. is “the core technology of new technological revolution and industrial transformation, is transcending the traditional means of simulating human intelligence by a computer.” The authors specifically focus on describing AI 2.0 and its application within the financial market because of the particular goals and suitability of this technology for the needs of the financial markets. Coeckelbergh and Reijers (2015, p. 172) state that “technologies have a temporal and narrative character: that they are made sense of by means of individual and collective narratives but also themselves co-constitute those narratives and inter-human and social relations; configuring events in a meaningful temporal whole.” The third stage of Fintech development implies the integration of big data with other elements such as the Internet and Blockchain for the achievement of better efficiency within the financial market. One example of an application that describes the efficiency of using A.I. in finance is the assessment of an individual’s credit score.
Next, in order to understand Blockchain, one must have sufficient knowledge of information centralization on the Internet. Zheng et al. (2018, p. 2) state that various barriers obstructing individuals from free data sharing existed until Blockchain was introduced, which is “a distributed, publicly available, and immutable ledger.” Blockchain and bitcoin cryptocurrency are inseparable as they were simultaneously introduced in 2008 by Nakamoto (DeVries 2016). In his paper, DeVries (2016) argues that cryptocurrencies, for example, bitcoin, do not have the potential of fully replacing the existing financial structures. However, they can transform the perception of market interactions. The conclusion made by the author suggests that due to the fact that cryptocurrencies imply fewer barriers or regulators, they can impact the existing approaches to currency exchange and national currencies perception.
Defining the concept of these currencies is crucial for understanding their application. DeVries (2016, p. 1) describes cryptocurrency as “an encrypted, peer-to-peer network for facilitating digital barter.” The author states that the primary advantage of this exchange model is the lack of oversight from a third party, which allows Internet users to exchange value more easily. According to Jani (2018, p. 1), “as of March 18, 2018, there are 1564 Cryptocurrencies available & traded in about 9422 exchanges.” Therefore, the current market of cryptocurrencies is large and will continue to grow to provide investment opportunities for individuals. Miraz and Ali (2018) focus on the prospects of blockchain technology and its prospective applications beyond the cryptocurrencies described in this paper. According to the authors, “distributed storage systems, proof-of-location, healthcare, decentralized voting, and so forth” are the prospective fields that can benefit from this technology in the future (Miraz & Ali 2018, p. 1). Therefore, Blockchain is not only applied in peer-to-peer value enhancement through cryptocurrencies but can be used in other fields as well.
A different perspective on the matter can help further improve the understanding of cryptocurrencies. Farell (2015, p. 130) provides the following description for the concept of cryptocurrencies – “virtual coinage system that functions much like a standard currency, enabling users to provides virtual payment for goods and services free of a central trusted authority.” The author argues that although the successful implementation of Blockchain occurred recently, the concept of cryptocurrencies was first explored in the 1980s.
The process of obtaining currencies is complex and requires one to solve an algorithm. Narayanan et al. (2016) describe mining puzzles that allow one to mine bitcoin and thus receive coins for the efforts. This is an essential element of the cryptocurrency market because, as was previously mentioned, the difficulty of mining is one of the aspects that affect the market. Thus, Narayanan et al. (2016) hypothesize that individuals will try to locate shortcuts for solving puzzles, which will allow them to achieve higher monetary rewards. This article provides a better understanding of the forces impacting the cryptocurrency market as well as implications for developing A.I. and data mining algorithms for predictions regarding value fluctuations.
Another background information that can be useful for this research is the historic change of bitcoin evaluation. Gandal and Halaburda (2016) examined the price changes and factors impacting them in the early stages of cryptocurrency implementation. The findings suggest that despite bitcoin being the most well-known and the most initial cryptocurrency introduced to the market, its value it is currently smaller than that of others. The authors applied the winner-take-all framework to analyze bitcoin against other cryptocurrencies. The Analysis presented by the authors of this study suggests that further models and algorithms should not be based on the implications of the new price trends of cryptocurrencies because those show little impact on the future development of these coins.
The cryptocurrency market incorporates a variety of tokens or currencies. According to Lee, Guo, and Wang (2018), apart from the popular bitcoin, several other cryptocurrencies were developed based on the original blockchain technology, which is referred to as altcoins. Among the reasoning for these actions are the limited 21 million bitcoins and the high demand for electricity required to mine these coins. Additionally, these elements led to the increased interest in cryptocurrencies that affected the financial market and, subsequently, the perceived value of these coins.
In regards to the prospects of bitcoin, Blockchain, and cryptocurrencies, a variety of suggestions is offered by scholars. For instance, Darlington (2014) states that cryptocurrencies, such as bitcoin or altcoins, have a specific meaning for developing economies because they mitigate the issue of inflation. However, limitations in regards to access and applicability still exist, as well as the possibility of uncovering problems with the mining algorithm in the future. Peters, Panayi, and Chapelle (2015) provide their insight on the topic of cryptocurrencies and prospective development trends, while ElBahrawy et al. (2017) state that even though new cryptocurrencies emerge and leave the market continuously, the overall development trend has remained stable since 2013. This is consistent with the neutral evolution model that accounts for elements that facilitate the development of this financial market.
It is critical to understand the specifics of operations that distinguish bitcoin from other cryptocurrencies. Delmolino et al. (2015) provide an understanding of smart contracts that were applied by altcoins emerging after bitcoin. Those use specific user-generated rules for transactions resulting in the ability to change the approaches to mining and currency exchange introduced by bitcoin. Thus, this article provides valuable insight into the issue of smart contracts, which can be used to enhance one’s knowledge of blockchain interactions.
Challenges and Limitations
Despite the overall benefits of cryptocurrencies, there are important issues that should be considered with their application and use. In his article, DeVries (2016) argues that contemporary in-state and international institutions are not tailored to the requirements of blockchain technology. Therefore, no clear regulations and legislation exist, which is both a challenge due to security and safety and an advantage due to the mitigation of unnecessary oversight. Azouvi, Maller, and Meiklejohn (2019, p. 127) state that “in a decentralized system, no one entity can act to censor transactions or prevent individuals from joining the network.” Another issue that can significantly affect this financial market and AI-based predictions is the need for user acceptance. DeVries (2016) argues that this is the only force that determines the success and value of any cryptocurrency; thus, with a change in people’s perception, the price of any cryptocurrency can change significantly.
The actual price of specific cryptocurrencies is unknown due to the current popularity of this technology. Salvetti, de Rossi, and Abbatemarco (2018) argue that due to the current demand and discussion surrounding Blockchain and cryptocurrencies, the current market is subjected to bias and hype, which does not allow one to adequately apply A.I. or data mining for Analysis. Many companies and venture capitalists invested in enterprises in this market, and the actual value of each cryptocurrency, including the most popular one, bitcoin, will be seen over time. Lindman, Rossi, and Tuunainen (2017) point out the limitation of Blockchain – is the network effect that impacts the cryptocurrencies and payment system limitations. More specifically, this technology is especially crucial for the financial markets because it eliminates a variety of risks associated with the industry. Additionally, Blockchain can be applied in many governmental operations, including issuing of certificates, which will allow additional accuracy and transparency to this process.
Among the specific limitations that exist, one should note that many countries, for instance, the U.S., introduced legislation aimed at regulating the cryptocurrency market. This can lead to a significant impact on the overall perception of the demand for coins. Hughes (2017) states that despite the current efforts, the decentralized nature of blockchain technology makes it impossible to enforce specific regulations. Delgado-Segura et al. (2018) state that P2P networks are the primary feature of cryptocurrencies that distinguishes them from other digital money. However, the issue is that no specific standard exists, which results in significant differences within the functionality of these networks.
How A.I., Cryptocurrency, and Blockchain are Related
The following paragraph will explore the connection between the three technologies. Hassani, Huang, and Silva (2018) state that currently, the cryptocurrency market is valued at trillions of U.S. dollars, which showcases its significance in the context of global finance. The authors argue that the primary connection between A.I. and cryptocurrencies is the fact that the latter requires the application of Big Data analysis due to its complexity and incorporation of a large number of users. This corresponds with the five features of big data – volume, variety, velocity, veracity, and value.
The technology of Blockchain is revolutionizing many elements of contemporary life, including the financial markets. Salvetti, de Rossi, and Abbatemarco (2018) state that within this model, users have a critical role because they need to continuously participate in the data exchange process to ensure the functioning of Blockchain. This allows individuals to be a part of the peer to peer exchanges or transactions. In this regard, it should be noted that the connection between Blockchain and bitcoin is the fact that the former was the first known successful application of blockchain technology. Therefore, cryptocurrencies emerged due to the development of the technology in question, and their further development, as well as the functioning of this financial market, will depend on this concept.
In regards to the application of A.I. and data mining, prior research suggests that information gathered via the Internet can be used to make valid predictions regarding the changes in the financial markets. Colianni, Rosales, and Signorotti (2015) present an example of Twitter and the gathering of information from this social media platform that helped researchers make predictions regarding the movement of securities. Subsequently, the authors argue that social media can be used to gather information about cryptocurrencies that will enable one to develop adequate trading strategies. Thus, one can say that based on the findings of the study by Colianni, Rosales, and Signorotti (2015, p. 1), a conclusion regarding the successful use of machine learning within the financial market of cryptocurrencies predictions can be made, more specifically, the following algorithms were tested by the authors – “logistic regression, Naive Bayes, and support vector machines.” The accuracy of such predictions was estimated at over ninety percent, which provides implications for the further development of similar strategies.
Cryptocurrencies are inevitably connected to Blockchain as they are a result of a block exchange. Thus, tokens that are received as part of this process have the potential of transforming the economic system globally because they allow one to exchange value with others independently. Laskowski and Kim (2016) explore the technology of text mining in the context of cryptocurrencies. The authors created a framework that incorporates a number of information streams from social media and messaging applications, which allowed them to develop a cohesive model representing the current financial market of cryptocurrencies. Therefore, unarguably, the nature of cryptocurrencies implies a need for a different analysis of information to make accurate predictions of market changes.
The Emergence of Cryptocurrency and Blockchains
As was previously mentioned, most researchers reviewed in this paper agree that the first cryptocurrency introduced was bitcoin. It was created in 2008, and according to DeVries (2016), to ensure parity, there is a specified number of this currency that can be generated by users. An unusual element of blockchain finance is explored by the author, who points out that, unlike traditional currencies, cryptocurrencies exist due to the perceived value that individuals are accepting them as payments place on this technology. For instance, a vendor receiving bitcoins has to believe that this currency has value since no institution can provide support for it. Due to this reason, one can argue that the blockchain finance market is more complicated when compared to traditional ones and requires a more advanced technology such as A.I. for proper Analysis and predictions.
Types of Blockchain
In order to understand the different kinds of Blockchain, one must have sufficient knowledge of the mechanisms that underline blockchain technology. Price (2017, p. 2) states that “a blockchain database is a distributed ledger comprising transactions and blocks.” This structure is critical for ensuring the safety of data because each block incorporates a hash from a previous block. Thus, it is highly unlikely that such data will be corrupted due to the integrity between different blocks. The cryptocurrencies that are created as a result of such exchanges or through the process called mining are intended to incentivize users to use the system and thus enable the exchange of information between different parties.
It should be noted that a variety of blockchain types exists that differ in accordance with the use of the algorithm. For instance, public blockchains use the proof of work model that can be seen in bitcoin. Users have to verify each transaction in order for this system to work. This model is open-sourced and, therefore can be used by anyone. Private blockchains, on the contrary, have limited access to data and thus can be used by a specified number of users.
Cryptocurrency Market Analysis
The formation of value within the cryptocurrency market is an essential feature that shapes the demand and price of each coin. Hayes (2015, p. 1308) argues that based on the empirical data and cross-sectional Analysis, it can be concluded that “the difficulty in ‘mining ‘for coins; the rate of unit production; and the cryptographic algorithm employed” are the critical elements. Kaplan, Aslan, and Bulbul (2018) focused on examining the word of mouth effect established through another social media platform Twitter and its prospective impact on the price of cryptocurrencies. The regression analysis revealed that a correlation between rumors regarding altcoins and price value exists. Yilmaz and Hazar (2018a) state that investors choose cryptocurrencies based on five primary factors, which allows one to create a cryptocurrency that would correspond to all these elements and make predictions regarding the market success of a particular coin. DeVries (2016, p. 2) argues that “cryptocurrencies could possibly be the single most disruptive technology to global financial and economic systems.” Thus, this technology is essential for the global economy and should be analyzed carefully.
Firstly, the decentralized nature of Blockchain mitigates the traditional disadvantages of other online payment methods – including commissions, chargebacks, risks of doubles endings, or possible fraud. Heid (2014) provides an assessment of this financial market and argues that Bitcoin is a successful proof of concept for blockchain technology. Despite the fact that the cryptocurrency market is more secure when compared to traditional financial markets, it has been subjected to various attackers. Heid (2014) provides examples of data breaches and attempts to target end-users and other malicious actions that were possible due to the fact that the protocol was experimental. The author presents the following explanation of the algorithm that allows blockchain transactions to function – the encrypted algorithm generates precomputed files, each containing a pair of public and private keys and assigned to a specific owner. Next, individuals engaged in transactions can send data stored in the file walled.dat on their hard drives to other users. Dynamic wallet addresses contain information about the private keys, while public keys contain information about the destination of a payment.
One crucial element of the cryptocurrency market is that a specified number of coins for each cryptocurrency exists. Therefore, the value of these elements is based on the supply and demand laws, which are significant forces within the cryptocurrency market. Zheng et al. (2018) conducted a study examining the existing articles that research the financial market and application of A.I. The authors introduced the concept of financial intelligence, which is critical for this market. In addition, the level of difficulty in regards to mining a coin impacts the final price because it fluctuates depending on the conditions. The nature of this protocol implies no need for a third party because the open-source protocol allows the transactions to be secure and reliable. In general, cryptocurrencies can be purchased through exchange marketplaces using fiat currencies. Thus one can conclude that “the market is diverse and provides investors with many different products” (Alessandretti et al. 2018, p. 1). Thus, forecasting within this domain will become increasingly important in the future.
To understand the specifics of the markets, a study focusing on their functioning was explored. Hitam and Ismail (n.d.) compare the performance of different machine learning algorithms in regards to their ability for prediction changes in the cryptocurrency market. The study uses technical analysis strategies for time series data forecasting, which implies an assessment of information within the market such as price, the volume of sales, and future predictions. It disregards other elements applied in the fundamental Analysis, including outside forces that may impact the financial markets.
In general, the algorithms proposed by the authors of the discussed study are reasonably successful in making accurate cryptocurrency market predictions. An article by Farell (2015) offers a comprehensive analysis of the cryptocurrency market, which provides an understanding of the underlying forces guiding its development. Yamada and Nakajima (2016) introduce the concept of micro pricing in Blockchain to develop a framework for understanding human behavior and its implications for the economy and financial markets. Gandal and Halaburda (2014) present an article that examines Competition in the cryptocurrency marker using the network effect model. The findings suggest that currently, cryptocurrencies are viewed as financial assets, which provides implications for developing A.I. for market predictions.
Algorithms and Methods of Predictions
Application of A.I. in the context of Analysis and prediction of the price is a valid strategy that can be used to make investment recommendations in regards to cryptocurrencies. However, many components should be considered. Another study that applies A.I. in the form of neural networks for the prediction of price fluctuation of cryptocurrencies was conducted by Gullapalli (2016), who used this framework to make predictions regarding the high and closing prices of bitcoin on a daily basis. Both time delay and recurrent neural networks were used in this experiment to account for a variety of factors that may impact the market, in order to train these neural networks, historical data regarding the price of bitcoin as a reference. Components such as quarterly highest and the lowest value, closing costs, and volume of demand were taken into account by Gullapalli (2016). In general, the algorithm developed by the author, more specifically the time-delay model, was successful at predicting the price.
The ethical element of A.I. analysis is an essential focus of this paper. Wallach (2010) explores the question of ethics and decision-making, the process that impacts the way humans choose to act. This research is relevant to the question of A.I. analysis applied to cryptocurrencies because, as was previously mentioned, this market is based solely on the value perception that individuals have. Cognitive mechanisms, such as reasoning, have an impact on this process, and with an appropriate model, A.I. can mirror this reasoning process, which will help make adequate predictions regarding cryptocurrencies. In his study, Wallac (2010) mainly focuses on developing robots that can be applied in a variety of domains without posing a threat by applying algorithms that allow them to use the ethical values of humans.
The emergence of Blockchain and cryptocurrencies is a result of a need for improvement in some aspects of financial operations. Jani (2018) explores the development and use of cryptocurrencies in India to provide an understanding of the enhancements that the introduction of this technology has, as well as the challenges that it poses. According to the author, approximately 21 countries responded to the widespread use of cryptocurrencies by introducing regulations aimed at protecting citizens from fraud. This affects the expectations of users and thus the ability to apply the A.I. algorithm for analyzing this market.
One can argue that the application of theories from traditional finance can be used to improve the A.I. algorithm in regard to cryptocurrencies. Khuntia and Pattanayak (2018) claim that the adaptive market hypothesis can be applied to cryptocurrencies and evaluate their theory using the example of bitcoin. In this regard, new information that people obtain has an impact on the price of their assets, in this case – cryptocurrencies, in accordance with the martingale difference sequence. However, Khuntia and Pattanayak (2018) state that over time, financial markets change and adapt to new conditions and are subjected to behavioral bias, which may be the primary limitation of data mining and A.I. analysis.
In order to prove this concept, several studies were examined that focus on behavior. Krafft, Penna, and Pentland (2018) conducted an experiment that can provide important implications for the future development of A.I. and data mining algorithms tailored to the Analysis of cryptocurrencies. The authors acknowledge the fluctuations of this market that are facilitated by bias and perception of the buyers and sellers and aim to account for it in their study. Krafft, Penna, and Pentland (2018) created bots that purchased small amounts of cryptocurrencies over a timeframe of six months to measure the impact of these actions on the overall market. The findings suggest that such actions have a significant short-term effect, which can be used as an important element of algorithm design.
Several researchers examined the prospects of machine learning and the plausibility of applying this strategy to the cryptocurrency market. Mini et al. (n.d., p. 96) studied the effect of applying neural networks on the predictions made regarding bitcoin prices using “multilayer perceptron (MLP) and Long short-term memory (LSTM) neural networks.” The authors added social and time elements to the standard model to improve precision. It was concluded that LSTM is more efficient due to the fact that it considers more factors. It can be found that the inclusion of a variety of factors that have a direct impact on the financial market, including behavioral elements, can significantly increase the accuracy of the A.I. algorithm. Napiah (2018) applied hybridization machine learning to facilitate the process of predicting changes in the cryptocurrency market. Other researchers focused on used a variety of A.I. frameworks to design a model that will accurately predict price fluctuations, For instance, Catania, Grassi, and Ravazzolo (2018), McNally (2016), and Jiang and Liang (2016) used A.I. while Yilmaz and Hazar (2018b) used conjoint Analysis. This presents a variety of evidence suggesting that A.I. and data mining can be used to make accurate predictions for cryptocurrencies.
Most researchers focus on applying traditional financial models when analyzing the market in question, which leads to several difficulties. Alessandretti et al. (2018) conducted a study targeting the machine learning algorithms that can be used to determine the market changes of cryptocurrencies. The authors used gradient boosting decision trees as the primary strategy for predicting the value of cryptocurrencies. The findings are based on the Analysis of 1,681 currencies and suggest that similar models that apply A.I. can be used to produce a profit from cryptocurrencies. Currently, Alessandretti et al. (2018, p. 2) state that the following algorithms were used to analyze cryptocurrencies, more specifically bitcoin – “random forests, Bayesian neural network, long short-term memory neural network, and other algorithms.” In general, from 2013 till 2017, the cryptocurrency market remained stable, showcasing its long-term properties. Li et al. (n.d.) focused on determining the prospects of using deep learning strategies in the prediction of cryptocurrency market fluctuations. The Python Library Keras was used for the development of this model, and two neural networks were created for the purpose of this experiment.
Different optimizers were used to adjust these models, and the findings suggest that it is plausible to develop a working neural network model that makes successful predictions regarding the value of cryptocurrencies, in this case, bitcoin, over short periods of time. Both classification and regression problems were resolved by Lee et al. (2018). This study provides implications for further development of deep learning networks that can be used to predict value changes accurately, as the authors of the described experiment aimed at determining changes within 3%. Pelletier (2018) uses Aylien API to make predictions regarding bitcoin. However, the study occurred during the plateau period experienced by the cryptocurrency, which was at the peak of its popularity prior. The author argues that news regarding bitcoin can significantly affect its price and suggests using Natural Language Toolkit as an A.I. model for Analysis. However, Pelletier (2018) states that according to the results of his studies, such information impacts only individual users and has little effect on investors. Similar findings are introduced by Lamon, Nielsen, and Redondo (2017), who argue that both news and social media can be used to predict the price fluctuations of the cryptocurrency market. More specifically, the authors focus on the three popular currencies – bitcoin, litecoin, and etherum.
Phillips and Gorse (2017) improved the previously described algorithm that accounted for news and social media rumors. This was done by applying epidemic modeling that can be used to detect prospective cryptocurrency bubbles and thus avoid risky investments. Lee, Ulkuatam, Beling, and Scherer (2018) used inverse reinforcement learning together with agent-based modeling to design an algorithm that would allow one to make accurate predictions in the cryptocurrency market.
The changes in the price that are prevalent for the market in question require additional attention. Conrad, Custovic, and Ghysels (2018) examined the volatility of bitcoin over both short and long timeframes using the GARCH-MIDAS model. The findings suggest that, in general changes in the price of bitcoin do not correspond with the standard volatility frameworks which can be observed in financial markets. However, the global economy and events affecting it have a direct impact on bitcoin, which provides implications for further Analysis of occurrences affecting cryptocurrencies. Radityo, Munajat, and Budi (2017) used artificial neural networks, while Karasu, Hacioglu, and Atlan (2018) applied time series machine learning. The results provide an understanding of successful strategies that can be used to predicts volatility.
Accounting for social media impact is critical for accurate cryptocurrency forecasting. Islam et al. (2018) used text mining for financial Analysis and understanding the implication of news in regards to market value. The study focuses on developing a variety of frameworks that can be used to classify significant information, which can be used for improving the A.I. algorithm. The authors suggest that “within the text mining techniques, predictive stock model (SPM) need to improve by the rank search method with the inclusion of gains and ratios along with forwarding selection methods by integrating dimensionality reduction techniques” (Islam et al. 2018, p. 770).
Cocco, Concas, and Marchesi (2017) conducted an experiment in which the researchers used an artificial model of the cryptocurrency market. This model is helpful due to its recumbence with the actual cryptocurrency market and ability to showcase autocorrelation. This study is useful because the model can be used to test different strategies of investment in the cryptocurrency market. Radityo, Munajat, and Budi (2017) conducted a study comparing several A.I. algorithms in their ability to predict the fluctuations of bitcoin accurately. The authors argue that A.I. was proven to have higher accuracy and efficiency than other strategies.
Additionally, one should note that features are included in the process of designing an A.I. network. Thierer and Castillo (2016) state that A.I. technology raises a variety of concerns from policymakers; however, the prospects of this technology can bring enormous economic and social benefits. When designing data mining and A.I. algorithms, some elements from the traditional financial analysis can be used to understand the basics of trading. Jegadeesh and Titman (1993) provide an assessment of the winners and losers framework that can be applied to the modern-day blockchain. This assessment offers implications for understanding the traditional financial market and buy-sell strategies that can be profitable.
Conclusion
Overall, this paper explores the topic of A.I. and the application of the algorithm in the Blockchain. In general, the findings suggest that the emergence of AI 2.0 and the development of the third Fintech stage can offer a variety of advantages for the financial industry. The studies examined in this paper suggest that A.I. and data mining can be applied to making an accurate prediction in regards to fluctuations in the financial market.
Reference List
Alessandretti, L, ElBahrawy, A, Aiello, LM, & Baronchelli, 2018, ‘Anticipating cryptocurrencies prices using machine learning’, Complexity, vol. 2018, pp. 1-16.
Azouvi, S, Maller, M & Meiklejohn, S 2019 ‘Egalitarian society or benevolent dictatorship: the state of cryptocurrency governance’, International Conference on Financial Cryptography and Data Security, Nieuwpoort, Curaçao, pp. 127-143.
Catania, L, Grassi, S & Ravazzolo, F 2018, Predicting the volatility of cryptocurrency time–series, Web.
Cocco, L, Concas, G, & Marchesi, M 2017, ‘Using an artificial financial market for studying a cryptocurrency market’, Journal of Economic Interaction and Coordination, vol. 12, no. 2, pp. 345–365
Coeckelbergh, M & Reijers, W 2015, ‘Cryptocurrencies as narrative technologies’, ACM SIGCAS Computers and Society, vol. 45, no. 3, pp. 172-178.
Colianni, S, Rosales, S & Signorotti, M 2015, Algorithmic trading of cryptocurrency based on twitter sentiment analysis, Web.
Conrad, C, Custovic, A & Ghysels, E 2018 ‘Long- and short-term cryptocurrency volatility components: a GARCH-MIDAS analysis’ Journal of Risk and Financial Management, vol. 11, no. 2, pp. 1-12.
Darlington, JK 2014, The future of bitcoin: mapping the global adoption of world’s largest cryptocurrency through benefit analysis, Web.
Delgado-Segura, S, Pérez-Solà, C, Herrera-Joancomartí, J, Navarro-Arribas, G & Borrell, J 2018, ‘Cryptocurrency networks: a new P2P paradigm’, Mobile Information Systems, vol. 3, pp. 1-16.
Delmolino, K, Arnett, M, Kosba, A, Miller, A, Shi, E 2016 ‘Step by step towards creating a safe smart contract: lessons and insights from a cryptocurrency lab’, Web.
DeVries, P.D. 2016, ‘An analysis of cryptocurrency, bitcoin, and the future’, International Journal of Business Managnment and Commerce, vol. 1, no. 2, pp. 1-9.
ElBahrawy, A, Alessandretti, L, Kandler, A, Pastor-Satorras, R, & Baronchelli A 2017 ‘Evolutionary dynamics of the cryptocurrency market’, Royal Society Open Science, vol. 4, no. 170623, pp. 2-9.
Farell, R 2015 ‘An analysis of the cryptocurrency industry’, Wharton Research Scholars, p. 130-152.
Gandal, N & Halaburda, H 2014,Competition in the cryptocurrency market, Web.
Gandal, N, & Halaburda, H 2016, ‘Can we predict the winner in a market with network effects? Competition in cryptocurrency market’, Games, vol. 16, no. 2, pp. 2-21.
Gullapalli, S 2004, Learning to predict cryptocurrency price using artificial neural network models of time series, Web.
Hassani, H, Huang, X & Silva, E 2018, ‘Big-crypto: big data, blockchain and cryptocurrency’, Big Data and Cognitive Computing, vol. 2, no. 4, p.34-49.
Hayes, A, 2015, ‘Cryptocurrency value formation: an empirical analysis leading to a cost of production model for valuing bitcoin’, MCIS 2015 Proceedings, Stockholm, Sweden, p. 1308-1321.
Heid, A 2014, Analysis of the cryptocurrency marketplace, Web.
Hitam, NA & Ismall, AR n.d. ‘Comparative performance of machine learning algorithms for cryptocurrency forecasting’, Indonesian Journal of Electrical Engineering and Computer Science, vol. 11, no. 3, pp. 1-11.
Hughes, SD 2017, Cryptocurrency regulations and enforcement in the U.S., Web.
Islam, MR, Al-Shailkhli, IM, Nor, R.B. & Varadarajan, V 2018, ‘Technical approach in text mining for stock market prediction: a systematic review’, Indonesian Journal of Electrical Engineering and Computer Science, vol. 10, no. 2, pp. 770-777.
Jani, S 2018, The growth of cryptocurrency in India: its challenges & potential impacts on legislation, Web.
Jegadeesh, N & Titman, S 1993, ‘Returns to buying winners and selling losers: implications for stock market efficiency’, The Journal of Finance, vol. 48, no. 1. pp. 65-91.
Jiang, Z & Liang, J 2016 ‘Cryptocurrency portfolio management with deep reinforcement learning’ Intelligent Systems Conference, London, UK, pp. 1-9.
Kaplan, C, Aslan, C & Bulbul, A 2018 ‘Cryptocurrency Word-of-Mouth Analysis viaTwitter’, Conference: II. International Conference on Theoretical and Applied Computer Science and Engineering 2018-Summer, Istanbul, Turkey, pp. 1-10.
Karasu, S, Hacioglu, R & Atlan, A 2018, ‘Prediction of Bitcoin prices with machine learning methods using time series data’, 2018 26th Signal Processing and Communications Applications Conference (SIU), Izmir, Turkey, pp. 1-4.
Khuntia, S, Pattanayak JK 2018, ‘Adaptive market hypothesis and evolving predictability of bitcoin’, Economics Letters, vol. 167, pp. 26-28.
Krafft, P, Penna, N & Pentland, A 2018 ‘An experimental study of cryptocurrency market dynamics’, Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems – CHI ’18, New York, USA, pp. 1-13.
Lamon, C, Nielsen, E & Redondo, EF 2017, Cryptocurrency price prediction using news and social media sentiment, Web.
Laskowski, M & Kim, H 2016, ‘Rapid prototyping of a text mining application for cryptocurrency market intelligence’, 2016 IEEE 17th International Conference on Information Reuse and Integration (IRI), Pittsburg, USA, pp. 1-6.
Lee, DK, Guo, L & Wang, Y 2018 ‘Cryptocurrency: a new investment opportunity?’, Journal of Alternative Investments, vol. 20, no. 3, pp. 16-40.
Lee, K, Ulkuatam, S, Beling, P & Scherer, W 2018 ‘Generating synthetic bitcoin transactions and predicting market price movement via inverse reinforcement learning and agent-based modeling’, Journal of Artificial Societies and Social Simulation, vol. 21, no. 3, p. 5-16.
Li, B, Timaul, D, Lukanan, A, Lau, J, Steinmetz, E n.d. Applying deep learning to better predict cryptocurrency trends, Web.
Lindman, J, Rossi, M & Tuunainen, V 2017, ‘Opportunities and risks of Blockchain Technologies in payments – a research agenda’, Proceedings of the 50th Hawaii International Conference on System Sciences, Hawaii, USA, pp. 1533-1542.
McNally, S 2016, Predicting the price of Bitcoin using Machine Learning, Web.
Miraz, MH & Ali, M 2018, ‘Applications of blockchain technology beyond cryptocurrency’, Annals of Emerging Technologies in Computing (AETiC), vol. 2, no. 1, pp. 1-6.
Misnik, S, Krutalevich, S, Prakapenka, P, Borovykh, P & Vasiliev M n.d., Neural network approximation precision change analysis on cryptocurrency price prediction, Web.
Napiah, N 2018, Prediction cryptocurrencies using hybridization machine learning, Web.
Narayanan, A, Bonneau, J, Felten, E, Miller, E & Goldfeder, S 2016, Bitcoin and cryptocurrency technologies: comprehensive introduction, Princeton University Press, Princeton.
Pelletier, C 2018, Computational approaches to predicting cryptocurrency prices, Web.
Peters, G, Panayi, E & Chapelle, A 2015, ‘Trends in crypto-currencies and blockchain technologies: A monetary theory and regulation perspective’, Journal of Financial Perspectives, vol. 3, no. 3, pp. 1-25.
Phillips, RC Gorse, D 2017, ‘Predicting cryptocurrency price bubbles using social media data and epidemic modelling’, 2017 IEEE Symposium Series on Computational Intelligence (SSCI), Honolulu, USA, pp. 1-7.
Price, T 2017, Predictive cryptocurrency mining and staking, Web.
Radityo, A, Munajat, O, Budi, I 2017, ‘Prediction of Bitcoin exchange rate to American dollar using artificial neural network methods’, 2017 International Conference on Advanced Computer Science and Information Systems (ICACSIS), Bali, Indonesia, pp. 1-5.
Salviotti, G, de Rossi, LM & Abbatemarco, N 2018, ‘A structured framework to assess the business application landscape of blockchain technologies’, Distributed Ledger Technology, the Blockchain, pp. 3467-3476.
Thierer, A & Castillo, A 2016, Preparing for the future of artificial intelligence, Web.
Wallach, W 2010 ‘Robot minds and human ethics: the need for a comprehensive model of moral decision making’, Ethics and Information Technology, vol. 12, no. 3, pp. 243-250.
Yamada, Y & Nakajima, T 2016 ‘Blockchain-LI: a study on implementing activity-based micro-pricing using cryptocurrency technologies’, MoMM ’16 Proceedings of the 14th International Conference on Advances in Mobile Computing and Multi Media, Singapore, Singapore pp. 203-207.
Yilmaz, NK & Hazar, H.B. 2018a, ‘Determining the factors affecting investors decision-making process in cryptocurrency investments’, Istanbul Finance Conference, Istanbul, Turkey, pp. 5-8.
Yilmaz, NK & Hazar, H.B. 2018b, ‘Predicting future cryptocurrency investment trends by conjoint analysis’, Finance and Accounting (JEFA), vol. 5, no. 4, pp.321-330
Zheng, X, Zhu, M, Li, Q, Chen, C & Tan, Y 2018 ‘FinBrain: when finance meets AI 2.0.’, Frontiers of Information Technology & Electronic Engineering, pp. 1-11.
