Applying Artificial Intelligence in Cryptocurrency Markets: A Survey
Round 1
Reviewer 1 Report
The paper presents overview of Machine Learning techniques used in Blockchain technology. The literature is well summarised and presented. Overall, the paper can be interest to wider community.
Some typos, which can be removed by careful reading.
Author Response
Reviewer 1 comment:
The paper presents overview of Machine Learning techniques used in Blockchain technology. The literature is well summarised and presented. Overall, the paper can be interest to wider community.
Some typos, which can be removed by careful reading.
Response:
We appreciate the reviewer for the positive comments regarding the paper. After careful reading, we have rectified a number of grammatical inconsistencies and typos.
Reviewer 2 Report
This study reviewed current research direction on current on cryptocurrencies. Some interesting insights were provided. The data collection method was well designed and systematic. The work provides useful reference to scholars and practitioners in the field. In addition, given FinTech is an emerging topic, few sentences could be added to add value to this work by mentioning the cryptocurrencies and FinTech in conclusion section as below.
“It’s worth mentioning that cryptocurrency is an important topic in FinTech (Chuen and Deng, 2017; Leong and Sung 2018, Stulz 2019), therefore, the findings from this study could also provide a reference for FinTech related studies.”
Relevant references are as below:
References:
Chuen, D. L. K., & Deng, R. H. (2017). Handbook of Blockchain, Digital Finance, and Inclusion: Cryptocurrency, FinTech, InsurTech, Regulation, ChinaTech, Mobile Security, and Distributed Ledger. Academic Press.
Leong, K., & Sung, A. (2018). FinTech (Financial Technology): what is it and how to use technologies to create business value in fintech way?. International Journal of Innovation, Management and Technology, 9(2), 74-78.
Stulz, R. M. (2019). Fintech, bigtech, and the future of banks. Journal of Applied Corporate Finance, 31(4), 86-97.
Author Response
Reviewer 2 comment:
This study reviewed current research direction on current on cryptocurrencies. Some interesting insights were provided. The data collection method was well designed and systematic. The work provides useful reference to scholars and practitioners in the field. In addition, given FinTech is an emerging topic, few sentences could be added to add value to this work by mentioning the cryptocurrencies and FinTech in conclusion section as below.
“It’s worth mentioning that cryptocurrency is an important topic in FinTech (Chuen and Deng, 2017; Leong and Sung 2018, Stulz 2019), therefore, the findings from this study could also provide a reference for FinTech related studies.”
Relevant references are as below:
References:
Chuen, D. L. K., & Deng, R. H. (2017). Handbook of Blockchain, Digital Finance, and Inclusion: Cryptocurrency, FinTech, InsurTech, Regulation, ChinaTech, Mobile Security, and Distributed Ledger. Academic Press.
Leong, K., & Sung, A. (2018). FinTech (Financial Technology): what is it and how to use technologies to create business value in fintech way?. International Journal of Innovation, Management and Technology, 9(2), 74-78.
Stulz, R. M. (2019). Fintech, bigtech, and the future of banks. Journal of Applied Corporate Finance, 31(4), 86-97.
Response:
We thank the reviewer for their encouraging comments regarding the paper, and its potential value in the FinTech community. Regarding cryptocurrencies and FinTech, we have included the sentence on Page 2 Line 55, where we feel it fits in with the discussion:
“Further references on FinTech companies and their approaches to dealing with the challenges and complexities of cryptocurrencies can be found in Chuen and Deng (2017); Leong and Sung (2018); Stulz (2019).”
Reviewer 3 Report
The authors aim to review the AI applications in Cryptocurrency markets. They examined current machine learning research that is used for cryptocurrency price forecasting. This paper has acceptable English language including a few spelling and writing mistakes.
The authors aims to answer the following research questions.
1. What factors influence the price of cryptocurrencies?
2. What is the state-of-the-art in AI research in the domain of cryptocurrency price prediction?
3. What are current gaps in the literature that may be addressed by conducting future research?
I think the authors need to clarify some points about the research. What is the main purpose of this study: i. Price prediction, ii. AI applications, iii. Determinants of the price of cryptocurrencies or all of them together.
There are different statement in the paper for the main purpose of this study. The title says “Artificial Intelligence Applications”: it means that I can read the papers about some algorithms finance data applications and I can learn something about the process of these applications.
As a second, I can read in many sections that the authors want to determine what factors influence the cryptocurrency prices. If the aim is to determine these factors, I must read about more relevant research paper review in literature section about the cryptocurrency price determinants such as economic, social, financial, political and any other aspect for the clarification of the research problem.
If the authors aim to make a forecast for the price or price movement prediction for cryptocurrencies, they need to investigate the technical indicators. Because financial technical indicators are very important tools to explain the behavior of financial data and to forecast the movement or real value of it. When we review the last published papers in the finance literature (I do not want to suggest any study due to journal research citation suggestion policy), it is easy to understand the importance of the technical indicators. We can see the usage of these indicators not only in stock market data but also in cryptocurrencies. Technical indicators showed better performance in explaining the behavior of the stock market and cryptocurrency assets in terms of the last researches. It would be better if the authors may review the last financial forecasting ML paper based on these indicators.
Especially in the last decade, numerous study was published in this field, as presented in the paper (Fig. 1). The author claimed that they presented a review study to fill the gaps in the literature with their recommendation. Nevertheless, I could not find any novel approach or recommendation in this paper to improve the current researches. There are a few basic suggestions that are insufficient. This topic would be very interesting and useful for future studies if it presents a novel forecasting framework for cryptocurrency price forecasting. This paper presented a literature review based on AI applications in cryptocurrencies that are easy to find in any other published paper. Therefore, when I compared this paper with published ones, unfortunately I could not find any significant contribution to the literature.
When I examined the researches presented in the study to explain the applications of the algorithms, the authors used machine-learning algorithms; however, they opted for the word “Artificial intelligence” in the title. The main idea is based on ML, so there is no need to use AI in the title.
The authors claimed that they used a cross-disciplinary approach to discuss the price determinants of altcoins from a financial and economic perspective in the conclusion section. This led to a conflict with the title and some statements in certain sections, as mentioned before.
I have read and scrutinized this paper. I believe that this paper tried to review a very comprehensive aspect of the problem based on a comparative approach. The paper is organized well and has fine English proficiency. I could not find any novel approach or suggestion for the research problems. This paper may be reviewed for consideration in a different journal. According to my professional opinion and as of my critics and suggestion, it is not suitable and sufficient for the "Algorithms".
Author Response
Reviewer 3 comment#1:
I think the authors need to clarify some points about the research. What is the main purpose of this study: i. Price prediction, ii. AI applications, iii. Determinants of the price of cryptocurrencies or all of them together.
Response:
We thank the reviewer for this comment concerning the clarity of the main purpose of the paper. In response, we would like to emphasise that the paper is about surveying the application of AI in cryptocurrency markets. The spectrum of applications ranges from predicting the price to portfolio management. To construct a common ground for readers without sufficient background, we decided to provide fundamental knowledge including cryptocurrency markets, blockchain, and so forth. In addition, we have in Subsection 2.2 emphasised the impact of price determinants (as part of pre-processing and feature engineering step) in developing an accurate model for price prediction. To enable a robust AI solution for predicting cryptocurrency prices, gaining a better understanding of price determinants goes a long way towards solving the problem. To further clarify the focus of the paper, we have made two changes:
- Page 3 Line 74: “The primary purpose of this paper is to review recent studies on AI applications in cryptocurrencies as well.”
- Page 3 Line 79: “Therefore, to achieve the main purpose of our study, this survey aims to answer the following questions to fill the gap in the existing surveys of recent publications in AI and cryptocurrency price prediction.”
Reviewer 3 comment#2:
There are different statement in the paper for the main purpose of this study. The title says “Artificial Intelligence Applications”: it means that I can read the papers about some algorithms finance data applications and I can learn something about the process of these applications. As a second, I can read in many sections that the authors want to determine what factors influence the cryptocurrency prices. If the aim is to determine these factors, I must read about more relevant research paper review in literature section about the cryptocurrency price determinants such as economic, social, financial, political and any other aspect for the clarification of the research problem.
Response:
We thank the reviewer for pointing out this concern. The main purpose of the paper is to provide a survey on the application of AI in dealing with challenges in cryptocurrency markets. To this end, we have done the following:
- Introduced cryptocurrencies in Section 2.1,
- Discussed publications concerning important factors in determining cryptocurrency prices in Section 2.2,
- Surveyed AI applications in price prediction in Section 3,
- Reviewed the application of reinforcement learning in cryptocurrency in Section 3.2.
Moreover, based on your comments, we feel that the title of the paper can be misinterpreted, and hence, we have changed the title to:
“Applying Artificial Intelligence in Cryptocurrency Markets: A Survey.”
Reviewer 3 comment#3:
If the authors aim to make a forecast for the price or price movement prediction for cryptocurrencies, they need to investigate the technical indicators. Because financial technical indicators are very important tools to explain the behavior of financial data and to forecast the movement or real value of it. When we review the last published papers in the finance literature (I do not want to suggest any study due to journal research citation suggestion policy), it is easy to understand the importance of the technical indicators. We can see the usage of these indicators not only in stock market data but also in cryptocurrencies. Technical indicators showed better performance in explaining the behavior of the stock market and cryptocurrency assets in terms of the last researches. It would be better if the authors may review the last financial forecasting ML paper based on these indicators.
Response:
We thank the reviewer for the comment on the importance of technical indicators. The point about the impact of indicators in understanding the price movement of financial markets is indeed very important, and researchers aiming at developing accurate predictive models must consider them. In this paper, we provide a survey on the application of AI models in dealing with challenges in cryptocurrencies as they are becoming a considerable part of investors’ portfolios. There are several AI models applied to cryptocurrency data to obtain a better understanding of the emerging market of cryptocurrencies, and we provide a categorisation of the information on this topic by concentrating on the models, the way they are compared, target features, data sources, and so forth. We believe our contribution to the literature via this paper will help researchers in this field to find relevant studies use appropriate algorithmic settings and data attributes to compare their own findings with other similar studies. In addition, we provide a list of plausible cross-disciplinary research directions to further investigate the markets in obtaining a better understanding and possibly in extracting profitable trade strategies.
Reviewer 3 comment#4:
Especially in the last decade, numerous study was published in this field, as presented in the paper (Fig. 1). The author claimed that they presented a review study to fill the gaps in the literature with their recommendation. Nevertheless, I could not find any novel approach or recommendation in this paper to improve the current researches. There are a few basic suggestions that are insufficient. This topic would be very interesting and useful for future studies if it presents a novel forecasting framework for cryptocurrency price forecasting. This paper presented a literature review based on AI applications in cryptocurrencies that are easy to find in any other published paper. Therefore, when I compared this paper with published ones, unfortunately I could not find any significant contribution to the literature.
Response:
We thank the reviewer for this comment. We emphasise that the main purpose of the paper is to provide a survey on the application of AI models in cryptocurrency markets, identify the gaps in the literature, and propose research directions that can be expected to fill the gaps. To gain deeper knowledge of cryptocurrency markets as evidenced in the publications we have surveyed, we believe the following areas warrant investigation:
- Integration within cryptocurrencies or with other financial assets,
- Macroeconomic factors as the states in reinforcement learning,
- The impact of the COVID-19 pandemic on the crypto markets,
- Sentiment analysis,
- An in-depth analysis of altcoins,
- Extreme condition detection in cryptocurrencies.
The reasons behind the potential of these topics is detailed in the manuscript. Take for example Item 2 in the list above (Macroeconomic factors as the states in reinforcement learning), the concept of algorithmic trading, and using reinforcement learning in detecting profitable trade ideas, are gaining popularity in FinTech companies. Two recent publications “Survey on Algorithmic Trading Using Sentiment Analysis” (Bagate et al., 2023) and “Sentiment and knowledge based algorithmic trading with deep reinforcement learning” (Nan at al.,2022) provide evidence of this. As evidenced in these publications, identifying states in a reinforcement learning model is a challenging task. Hence, to construct a reliable automated trading platform, a proper way of incorporating microeconomic factors in the definition of states can be considered as a considerable contribution in the field of algorithmic trading of cryptocurrencies.
Reviewer 3 comment#5:
When I examined the researches presented in the study to explain the applications of the algorithms, the authors used machine-learning algorithms; however, they opted for the word “Artificial intelligence” in the title. The main idea is based on ML, so there is no need to use AI in the title.
Response:
We thank the reviewer for the comment on the title of the paper. The delineation between AI and ML is not obvious as mentioned in “Machine learning and deep learning” (Janiesch et al., 2021) and “Application of deep learning algorithms in geotechnical engineering: a short critical review” (Zhang et al., 2021), and the community currently has settled on a framework that considers AI as an umbrella topic, and ML sits within that umbrella as a subset of AI. In line with the ambiguity that the reviewer mentioned, we stated the following statement in the paper:
“There is no clear border to distinguish different topics in AI and ML. However, it is essential to differentiate these relevant concepts from each other for an adequate understanding of AI.”
In addition, based on the publications we reviewed in the paper, we have adapted a categorisation of the models in AI and ML as shown in Figure 4 in the manuscript.
Furthermore, in response to the reviewer’s comment, we made the following change on Page 9 Line 303 in the paper to remove the ambiguity:
“We follow the frequently used framework in categorising AI and ML models which considers ML as a subset of AI.”
Reviewer 3 comment#6:
The authors claimed that they used a cross-disciplinary approach to discuss the price determinants of altcoins from a financial and economic perspective in the conclusion section. This led to a conflict with the title and some statements in certain sections, as mentioned before.
Response:
We thank the reviewer for pointing out the conflict. The first sentence of the paragraph in the Conclusion needs clarification, and hence, we have modified this sentence to improve its clarity as follows.
“This paper first uses a cross-disciplinary approach to discuss the price determinants of cryptocurrencies from a financial and economic perspective. Then, recent studies on the use of various AI models in cryptocurrency price prediction are reviewed through a comparative survey.”
Round 2
Reviewer 3 Report
Dear authors,
Thank you for your response to my comments and your effort.
Sincerely
