**1. Introduction**

Cryptocurrency is a private system that enables trades between individuals without a central and intermediate agency. In early 2009, Bitcoin (BTC) was valued for the first time at US\$0.08. The currency fluctuated for more than four years until the price touched \$1110 in 2013. Due to high volatility and massive fluctuations in prices in cryptocurrencies, accurate price predictions are a complex and challenging task. That is mainly because the costs of

**Citation:** Erfanian, S.; Zhou, Y.; Razzaq, A.; Abbas, A.; Safeer, A.A.; Li, T. Predicting Bitcoin (BTC) Price in the Context of Economic Theories: A Machine Learning Approach. *Entropy* **2022**, *24*, 1487. https:// doi.org/10.3390/e24101487

Academic Editors: Stanisław Drozd˙ z,˙ Marcin W ˛atorek and Jarosław Kwapie ´n

Received: 24 August 2022 Accepted: 13 October 2022 Published: 18 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

cryptocurrency move unpredictably and chaotically. Machine learning techniques may help bring in some methodology that will lead to better solutions to the problem. In the last several years, there has been an increasing interest in using machine learning techniques in different areas of science [1,2], particularly cryptocurrency price forecasting [3]. For instance, Dutta et al. [4] used macroeconomic indicators, including interest rates, S&P500 market returns, US bond yields, and gold price level as predictive variables for daily BTC prices. The results show that macroeconomic indicators have short-term predictability power. Wang and Vergne [5] investigated macroeconomic indicators, namely supply growth, defined as BTCs in circulation, to see their effect on BTC return. They found that an increase in supply is positively associated with weekly returns. Conrad et al. [6] found that S&P500 volatility has a significantly positive effect on long-term BTC volatility.

Jang and Lee [7] investigated the effect of blockchain information, including average block size, miner revenue, mining difficulty, and hash rate, on BTC prices. Their results proved that the recent volatility in BTC prices stems from the blockchain information indicators. Wang and Vergne [5] investigated blockchain information indicators, including several unique collaborators contributing code to the project, the number of proposals merged in the core codebase, the number of issues raised by the community about the code, and fixed the developer's number of forks on BTC returns. They found a positive and significant relationship between blockchain information variables and weekly returns. Therefore, the first research question arises: (1) What are the significant variables as shortterm or long-term BTC price predictors? In addition, much previous research on BTC price predictions with machine learning is conducted either using machine learning techniques or conventional statistical analysis without enough theoretical and analytical support. This study investigates whether the macroeconomic, microeconomic, and blockchain information indicators based on economic theories predict the BTC price. According to these considerations, the second research question is: (2) What are the underlying economic theories of BTC price predictors?

There is not enough available literature on BTC price prediction on Google Scholar compared to stocks: around 400 papers about BTC price prediction problems with machine learning algorithms. There are almost 5500 papers about stock price prediction with machine learning algorithms. Also, according to the existing literature, some research on the BTC price prediction problem shows that machine learning outperforms conventional statistical analysis. At the same time, some still believe that traditional models can predict the BTC price better. For instance, Chen et al. [8] applied machine learning techniques models, including random forest, XGBoost, quadratic discriminant analysis, SVM, and LSTM, and statistical methods, including logistic regression and linear discriminant analysis, to predict high-frequency BTC price. They found that Statistical methods achieve an accuracy of 66%, outperforming more complicated machine learning algorithms for daily BTC price prediction. However, machine learning for BTC's 5-min interval price prediction is superior to statistical methods, with accuracy reaching 67.2%. Pang et al. [9] compared neural network models, sentiment data models, and conventional technical indicators and decision trees to predict BTC prices. The analysis found that the robust neural network models offer better accuracy in predicting BTC prices. Therefore, more research should show whether machine learning algorithms are superior to statistical analysis. Hence, the third research questions are: (3) Are machine learning algorithms superior to traditional methods for BTC price prediction? What machine learning model performs better? What are the best feature selection techniques?

The research innovation herein is looking at BTC price prediction through theoretical aspects. The overall findings show that SVR is superior to other machine learning models and traditional models. This paper has several contributions. It can contribute to international finance to be used as a reference for setting asset pricing and improved investment decision-making. It will be helpful for central bankers, traders, investors, and portfolio managers. Also, it contributes to the economics of BTC price prediction by introducing its theoretical background. Moreover, as the authors still doubt whether machine learning can

beat the traditional methods in BTC price prediction, this research contributes to machine learning configuration and helping developers to use it as a benchmark. The rest of the paper is as follows. In the literature review section, there is an overview of existing work and differences from the current work. After that, the methodologies used in this research are briefly explained. Subsequently, the results and discussion are presented. In the end, the paper is concluded.
