|
|
--- |
|
|
language: en |
|
|
license: mit |
|
|
tags: |
|
|
- sentiment-analysis |
|
|
- text-classification |
|
|
- yelp |
|
|
- transformers |
|
|
- distilbert |
|
|
--- |
|
|
|
|
|
# Yelp Reviews Sentiment Analyzer |
|
|
|
|
|
## Model Overview |
|
|
This is a DistilBERT-based sentiment analysis model fine-tuned on a subset of the Yelp Open Dataset. It classifies restaurant reviews into three categories: Negative, Neutral, and Positive. |
|
|
|
|
|
## Intended Use |
|
|
- Sentiment classification of restaurant reviews for business insights, customer feedback analysis, or academic research. |
|
|
- Can be integrated into applications to provide real-time sentiment detection. |
|
|
|
|
|
## Training Data |
|
|
- Yelp Open Dataset (restaurant reviews subset). |
|
|
- Labels derived from star ratings converted into sentiment classes. |
|
|
|
|
|
## Model Architecture |
|
|
- Based on `distilbert-base-uncased`. |
|
|
- Fine-tuned using Hugging Face's `AutoModelForSequenceClassification`. |
|
|
|
|
|
## Performance |
|
|
- Accuracy: ~78.5% |
|
|
- F1 Score: ~78.4% |
|
|
- Precision: ~78.3% |
|
|
- Recall: ~78.5% |
|
|
|
|
|
## Limitations |
|
|
- Performance may vary on reviews from domains outside Yelp restaurants. |
|
|
- Model is trained only on English-language reviews. |
|
|
- Neutral class can be subjective, and borderline cases may be misclassified. |
|
|
|
|
|
## How to Use |
|
|
Use Hugging Face Transformers pipeline: |
|
|
|
|
|
```python |
|
|
from transformers import pipeline |
|
|
|
|
|
sentiment_analyzer = pipeline("sentiment-analysis", model="fitsblb/YelpReviewsAnalyzer") |
|
|
|
|
|
result = sentiment_analyzer("The food was amazing but the service was slow.") |
|
|
print(result) |
|
|
|
