Machine Learning-Based Approach for Predicting the Altcoins Price Direction Change from a High-Frequency Data of Seven Years Based on Socio-Economic Factors, Bitcoin Prices, Twitter and News Sentiments

Anamika Gupta (), Gaurav Pandey (), Rajan Gupta (), Smaran Das (), Ajmera Prakash (), Kartik Garg () and Shreyan Sarkar ()
Additional contact information
Anamika Gupta: S.S. College of Business Studies, University of Delhi
Gaurav Pandey: IIT Jodhpur
Rajan Gupta: Analyttica Datalab
Smaran Das: S.S. College of Business Studies, University of Delhi
Ajmera Prakash: S.S. College of Business Studies, University of Delhi
Kartik Garg: S.S. College of Business Studies, University of Delhi
Shreyan Sarkar: S.S. College of Business Studies, University of Delhi

Computational Economics, 2024, vol. 64, issue 5, No 15, 3026 pages

Abstract: Abstract Altcoins are alternative types of coins under cryptocurrency, apart from traditional Bitcoins, for which predicting the price movement presents a multifaceted challenge deeply rooted in the volatile nature of the cryptocurrency market. This study compares and analyzes different Machine Learning (ML) and Deep Learning (DL) models for price movement prediction through diverse data sources like Bitcoin prices, social media sentiments, and news sentiments, apart from different socio-economic factors specific to USA geography due to its maturity on use of Altcoins, with temporal scope spanning from 2016 to 2022 collating over 77 M tweets and news items. Ethereum, Binance, XRP, Cardano, Monero, Tron, Stellar, and Litecoin, were considered for experimentation across widely used algorithms like Gradient Boosting, Naive Bayes, Decision Trees, Neural Networks, and the like, with different day-length lags ranging up to 4 days. Highly relevant features were selected using Random Forest selection method and highly correlated features have been removed before the modeling. Accuracy for price movement prediction models varied from 71.03% for Ethereum to 66.14% for Stellar, which were better by 15–20% as compared to percentage benchmarking done by literature to be ranging around 50 s and 60 s. For the model validation, sensitivity analysis involving day-wise lag analysis, and different data splits (based on size and months) were considered, which was stable for the high performing models. Further, an interesting result was observed during the study. In order of priority, Bitcoin prices, social media sentiments, and news sentiments significantly impact altcoin price movement. This implies that by studying the Bitcoin price movement and market sentiments, investors can make wise decisions towards altcoin investments. This study holds significance for researchers and practitioners to understand the impact in the trading market of cryptocurrency and help an investor diversify their portfolio. The findings will be helpful for Algo Trading Platforms, Financial Advisors, Trading Experts, Industry Experts, Researchers, and Scholars.

Keywords: High-frequency; Altcoins; Socio-economic factors; News sentiments; Machine learning; Twitter sentiments; Lag (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1007/s10614-023-10538-5 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:kap:compec:v:64:y:2024:i:5:d:10.1007_s10614-023-10538-5

Ordering information: This journal article can be ordered from
http://www.springer. ... ry/journal/10614/PS2

DOI: 10.1007/s10614-023-10538-5

Access Statistics for this article

Computational Economics is currently edited by Hans Amman

More articles in Computational Economics from Springer, Society for Computational Economics Contact information at EDIRC.
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().