Model Card for tripadvisor_review_rater

Model Details

Model Description

This model predicts a 1–5 star rating based on English-language hotel reviews, fine-tuned on the TripAdvisor Hotel Reviews dataset.

• Developed by: @MLandMarketing
• Model type: Text classification (5-class sentiment rating prediction)
• Language(s) (NLP): English
• License: MIT
• Finetuned from model: bert-base-uncased
• Repository: tripadvisor_review_rater

Uses

Direct Use

• Predict star ratings from hotel review text (1–5 stars)

Out-of-Scope Use

• Not suitable for non-English text
• Not recommended for commercial deployment without further validation

---

Bias, Risks, and Limitations

• Reflects biases in the original TripAdvisor review data (e.g., regional, cultural, linguistic)
• May struggle with sarcasm, slang, or subtle sentiment shifts

---

Recommendations

Users (both direct and downstream) should be made aware of the model’s limitations and potential biases. Further validation is advised before deployment in production or decision-making contexts.

---

How to Get Started with the Model

from transformers import AutoTokenizer, AutoModelForSequenceClassification

tokenizer = AutoTokenizer.from_pretrained("MLandMarketing/tripadvisor_review_rater")
model = AutoModelForSequenceClassification.from_pretrained("MLandMarketing/tripadvisor_review_rater")

review = "Room was clean and comfortable, but breakfast was disappointing."
inputs = tokenizer(review, return_tensors="pt", truncation=True, padding=True)
pred = model(**inputs).logits.argmax(dim=1).item()
print("Predicted Rating:", pred + 1)

Training Details

• Data: TripAdvisor Hotel Reviews (20,000+ reviews)
• Epochs: 3
• Batch size: 16
• Optimizer: AdamW
• Loss function: CrossEntropyLoss
• Final Train Loss: 0.64
• Accuracy: 67.7%
• Mean Absolute Error (MAE): 0.34 stars

Environmental Impact

• Hardware: Google Colab GPU (Tesla T4)
• Training time: ~1.3 hours
• Cloud Provider: Google
• Carbon Emitted: Minimal (small model, short training time)

Citation

BibTeX:

@misc{tripadvisorreviewrater2025,
  author       = {MLandMarketing},
  title        = {Tripadvisor Review Rater - BERT-based Review Rating Classifier},
  year         = {2025},
  howpublished = {\url{https://huggingface.co/MLandMarketing/tripadvisor_review_rater}}
}

APA:

MLandMarketing. (2025). Tripadvisor Review Rater - BERT-based Review Rating Classifier. Hugging Face. https://huggingface.co/MLandMarketing/tripadvisor_review_rater