Sentiment-Driven LSTM Analysis of Bitcoin Price: Uncovering Insights from Tweets and Macroeconomics Data

Understanding the multifaced factors that drive Bitcoin's price movements has been a challenging task. While past research has primarily investigated the correlation between macroeconomic indices and market prices, there is growing recognition that sentiment expressed on platforms like Twitter...

Full description

Saved in:
Bibliographic Details
Published in2024 IEEE 48th Annual Computers, Software, and Applications Conference (COMPSAC) pp. 285 - 290
Main Authors Hsu, Chia-Chun, Lu, Po-Han, Chu, Jyun-Siyan, Chang, Ya-Ning
Format Conference Proceeding
LanguageEnglish
Published IEEE 02.07.2024
Subjects
Bitcoin
Deep learning
LSTM
Macroeconomic factor
Online banking
Predictive models
Sentiment analysis
Social networking (online)
Online AccessGet full text

Cover

More Information
Summary:Understanding the multifaced factors that drive Bitcoin's price movements has been a challenging task. While past research has primarily investigated the correlation between macroeconomic indices and market prices, there is growing recognition that sentiment expressed on platforms like Twitter can also influence market behavior, particularly in the short term. Therefore, this study develops a price prediction system using sentiment analysis and deep learning techniques to explore the influence of both macroeconomic and sentiment indicators on the dynamics of Bitcoin's price. The results show that incorporating sentiment features can enhance the accuracy of the Bitcoin price prediction, particularly for the 1-week and 1-month time windows. However, when considering the longer time window, macroeconomic indices demonstrate superior performance.
ISSN:2836-3795
DOI:10.1109/COMPSAC61105.2024.00047