Publication

Twitter impact on Bitcoin price

View document

Bibliographic Details
Summary:Ten years ago, the Bitcoin price was $133 on September 30, 2013. On September 30, 2023, Bitcoin registered a price of over $27,000. Bitcoin is exchanged similarly to stocks in the stock market, operating independently of traditional financial systems for portfolio diversification. Twitter played a significant role in Bitcoin's rise, using social media platforms to spread information and build communities. These communities promoted Bitcoin, creating hype. The objective is to understand if tweets impact Bitcoin's price movements and if their analysis is useful for forecasting. CryptoBERT and VADER sentiment analysis tools will be used comparing their effectiveness. Various Machine Learning models such as Random Forest, XGBoost, AdaBoost, SVM, KNN, Bayesian Regression, and GBM will be applied for a comprehensive understanding. The analysis will focus on active social media users to determine the accuracy of their information in forecasting Bitcoin fluctuations. The study's findings may contribute to ongoing research on sentiment analysis for a broader range of financial instruments, including Bitcoin.
Main Authors:Jorge, Pedro Gaspar
Subject:Bitcoin Twitter Sentiment Analysis Vader CryptoBert Machine Learning
Year:2023
Country:Portugal
Document type:master thesis
Access type:open access
Associated institution:Universidade de Lisboa
Language:English
Origin:Repositório da Universidade de Lisboa
Description
Summary:Ten years ago, the Bitcoin price was $133 on September 30, 2013. On September 30, 2023, Bitcoin registered a price of over $27,000. Bitcoin is exchanged similarly to stocks in the stock market, operating independently of traditional financial systems for portfolio diversification. Twitter played a significant role in Bitcoin's rise, using social media platforms to spread information and build communities. These communities promoted Bitcoin, creating hype. The objective is to understand if tweets impact Bitcoin's price movements and if their analysis is useful for forecasting. CryptoBERT and VADER sentiment analysis tools will be used comparing their effectiveness. Various Machine Learning models such as Random Forest, XGBoost, AdaBoost, SVM, KNN, Bayesian Regression, and GBM will be applied for a comprehensive understanding. The analysis will focus on active social media users to determine the accuracy of their information in forecasting Bitcoin fluctuations. The study's findings may contribute to ongoing research on sentiment analysis for a broader range of financial instruments, including Bitcoin.