Financial Sentiment Analysis Model
Model Description
This model is a fine-tuned version of the distilbert-base-uncased transformer model, specifically trained to perform sentiment analysis on financial news articles. The model classifies the sentiment of the given text into three categories: positive, negative, and neutral.
Training Data
The model was trained on the Financial News Sentiment Dataset, which includes financial news articles annotated for sentiment analysis.
Model Training
Pre-trained Model
- Model:
distilbert-base-uncased - Tokenizer:
distilbert-base-uncased
Training Parameters
- Learning rates:
2e-5,3e-5 - Number of epochs:
3,4 - Batch size:
16 - Weight decay:
0.01
Best Hyperparameters
The best hyperparameters found during hyperparameter tuning were:
- Learning rate:
2e-5 - Number of epochs:
3
Performance
The model was evaluated using accuracy, precision, recall, and F1 score. The best accuracy achieved was `94.7%'.'
Usage
You can use this model for sentiment analysis on financial news articles by loading it with the Hugging Face Transformers library.
Model Limitations
- The model is specifically trained on financial news articles and may not generalize well to other domains.
- It might not capture the nuances of financial sentiment entirely due to the complexity of financial language.
Acknowledgments
This model was trained using the Hugging Face Transformers library and the financial news sentiment dataset by Jean-Baptiste. Special thanks to the creators of the dataset and the Hugging Face team.
- Downloads last month
- 103