Empirical Study of Cryptocurrency Prices Using Linear Regression Methods

With a market cap of 1.9 trillion and more than 10 thousands active trading cryptocurrencies, the global crypto market is claimed to be an attractive and vibrant market that strongly attracts many participants. Studying cryptocurrency price tendency is one of the most challenging and interesting res...

Full description

Saved in:
Bibliographic Details
Published in2022 RIVF International Conference on Computing and Communication Technologies (RIVF) pp. 701 - 706
Main Authors Van Tran, Loc, Le, Son Thanh, Tran, Ha Manh
Format Conference Proceeding
LanguageEnglish
Published IEEE 20.12.2022
Subjects
Bitcoin
Computational modeling
Correlation
Cryptocurrency Price
Data models
Lasso Regression
Measurement
Multiple Linear Regression
Prediction
Prediction algorithms
Predictive models
Ridge Regression
Online AccessGet full text

Cover

More Information
Summary:With a market cap of 1.9 trillion and more than 10 thousands active trading cryptocurrencies, the global crypto market is claimed to be an attractive and vibrant market that strongly attracts many participants. Studying cryptocurrency price tendency is one of the most challenging and interesting research fields. Despite the increasing number of studies tackling this field, it is essential to understand the factors influencing the price and analyze the most efficient model for working with crypto data. This study applies three regression models: Multiple Linear Regression, Ridge Regression, and Lasso Regression for cryptocurrency price prediction. By exploiting features that are directly tied to the closing price, the study evaluates the performance of three models on four distinct cryptocurrencies. The final results reveal the outstanding performance of Lasso Regression that could be applied in the crypto context for future studies.
DOI:10.1109/RIVF55975.2022.10013790