Crypto Sentiment Analyzer
Fine-tuned BERT model for cryptocurrency sentiment analysis from social media, news, and forum posts.
Model Details
- Base Model:
bert-base-uncased - Task: Text Classification (3 labels: NEGATIVE, POSITIVE, NEUTRAL)
- Training Data: 50K crypto tweets + Reddit posts (2024-2025)
- Accuracy: 92.3% on validation set