Can ChatGPT Predict Bitcoin Prices? Twitter Sentiment Analysis on Bitcoin Using ChatGPT

Abstract

Cryptocurrencies, particularly Bitcoin, have garnered significant attention in recent years. The in trinsic nature of Bitcoin requires a comprehensive understanding of investor sentiment to accurately predict its price movements. Traditionally, lexicon-based sentiment analysis models, such as VADER and TextBlob, have been widely used. However, their effectiveness in predicting price movements re mains limited. This paper explores an alternative approach by leveraging ChatGPT, an advanced large language model known for its ability to understand and generate natural language. ChatGPT's ability to compre hend human language nuances and context makes it a promising candidate for sentiment analysis. The study compares sentiment scores generated by ChatGPT with those from traditional lexicon based sentiment analysis models through analyzing their correlation with Bitcoin price movements over time. By examining both recent and historical data, we aim to identify patterns and assess the predictive power of sentiment-based models over different market conditions.