Database of twitter influencers in cryptocurrency (2021-2023) with sentiments

Area of Science:

• Computational Social Science
• Financial Technology (FinTech)
• Artificial Intelligence

Background:

• Social media platforms like Twitter host expert opinions (influencers) that significantly impact public perception.
• Cryptocurrency markets are volatile, necessitating advanced methods for trend prediction.
• Sentiment analysis of expert opinions offers a novel approach to understanding market dynamics.

Purpose of the Study:

• To investigate the predictive power of influencer sentiment on cryptocurrency price trends.
• To develop and apply a hybrid deep learning model for accurate sentiment analysis of financial tweets.
• To create a specialized dataset of influencer opinions on cryptocurrencies for trend analysis.

Main Methods:

• A hybrid deep learning model combining RoBERTa (Robustly Optimized BERT Pretraining Approach) and BiGRU (Bidirectional Gated Recurrent Unit) was employed for sentiment analysis.
• A unique dataset was curated, comprising tweets from over 52 influencers regarding eight cryptocurrencies, collected over eight months (February 2021 - June 2023).
• The dataset includes sentiment polarity, compound scores, importance coefficients, and historical price data for major cryptocurrencies like Bitcoin, Ethereum, and Binance.

Main Results:

• The study demonstrates that influencer sentiment, when analyzed with the hybrid model, correlates with cryptocurrency price movements.
• The specialized dataset allows for the determination of dominant daily sentiment polarity for specific cryptocurrencies.
• The RoBERTa-BiGRU model effectively captures nuanced opinions from tweets, distinguishing them from general hashtag-based sentiment.

Conclusions:

• Expert opinions shared on social media, when analyzed through advanced AI techniques, can serve as a valuable tool for predicting cryptocurrency market trends.
• The developed sentiment analysis framework provides actionable insights for investors seeking to navigate the cryptocurrency market.
• This research highlights the potential of integrating social media intelligence with financial data for enhanced market forecasting.