Sentiment Analysis Model
This model is a fine-tuned version of DistilBERT for sentiment analysis. It classifies text into three categories: Positive, Neutral, and Negative.
Model Details
- Model Type: DistilBERT (DistilBERT-base-uncased)
- Fine-Tuning Task: Sentiment Analysis
- Classes: 3 (Positive, Neutral, Negative)
- Dataset: Custom sentiment dataset with labeled "Positive", "Neutral", and "Negative" text data.
Intended Use
This model can be used to classify text as Positive, Neutral, or Negative. It's ideal for applications that require sentiment classification, such as customer feedback analysis, reviews, or social media sentiment monitoring.
Model Performance
This model was fine-tuned on a custom sentiment dataset. Below are its performance metrics on the validation dataset:
- Accuracy: 0.89
- F1-Score: 0.88 (Macro Avg)
- Precision: 0.89 (Weighted Avg)
- Recall: 0.89 (Weighted Avg)
Performance by Class:
- Positive: Precision: 0.88, Recall: 1.00, F1-Score: 0.93
- Neutral: Precision: 0.89, Recall: 0.74, F1-Score: 0.81
- Negative: Precision: 0.89, Recall: 0.92, F1-Score: 0.90
License
This model is licensed under the MIT License.
Usage
Install Hugging Face Transformers
To use this model, you need to install the transformers library. You can do so with the following command:
pip install transformers
Example Code to Use the Model
from transformers import pipeline
# Load the sentiment-analysis pipeline
sentiment_analysis = pipeline("text-classification", model="chawki17/my_sentiment_model")
# Example text
text = "I love this product!"
# Predict sentiment
result = sentiment_analysis(text)
print(result)
Inputs and Outputs
- Input: A string of text (e.g., a customer review).
- Output: A sentiment label (Positive, Neutral, or Negative) with confidence score.
Example output:
[{'label': 'POSITIVE', 'score': 0.98}]
Limitations
- The model may not perform well on text data from domains that were not part of the training set.
- It may not generalize well to very short texts or highly domain-specific language.
- The model is based on English text data and may not work well for other languages.
Model Card and Documentation
For more details on this model and its performance, visit the model page on Hugging Face.
Acknowledgements
- This model was fine-tuned using the DistilBERT architecture.
- The dataset was custom-built for this sentiment analysis task.