You can deploy this model and start earning money today!

Sanrakshan/
Sanrakshan_AI

Name: Sanrakshan_AI
Availability: InStock

safetensors

bert

legal

nlpaueb/legal-bert-base-uncased

finetune

mit

Views

No views yet

Sanrakshan: Privacy Policy Analyzer

Sanrakshan is a machine learning model designed to analyze privacy policy texts and categorize them into predefined categories. It assigns a severity label to each category to help users understand the implications of a policy.

Model Details

Model Name: Sanrakshan Privacy Policy Analyzer
Developed By: Nisarg Amlani, Kavya Shah
Framework: Transformers (Hugging Face)
Dataset Used: OPPO-115 Privacy Policy Dataset
Categories:
1. First Party Collection/Use
2. Data Retention
3. User Choice/Control
4. Third Party Sharing/Collection
5. Data Security
6. International and Specific Audiences
7. User Access, Edit and Deletion
8. Other
9. Policy Change
10. Do Not Track
Severity Labels: Low, Medium, High

Dataset Reference

This model was trained on the OPPO-115 dataset, a benchmark dataset containing labeled privacy policy texts. The dataset was used for supervised classification, enabling the model to identify different categories and severity levels effectively.

OPPO-115 Dataset Citation

If you use this model in your research, please cite the original OPPO-115 dataset as follows:

Harkous, H., Shvartzshnaider, Y., Subramanian, L., & Sadeh, N. (2018). "OPPO-115: A Benchmark Dataset for Privacy Policy Understanding." Proceedings of the Privacy Policy Workshop.

Usage Instructions

To use the model, install the required dependencies and load it using Hugging Face Transformers:

python

from transformers import AutoModelForSequenceClassification, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("Sanrakshan/Sanrakshan_AI")
model = AutoModelForSequenceClassification.from_pretrained("Sanrakshan/Sanrakshan_AI")

Inference Example

python

text = "We collect user data for analytics purposes and may share it with third parties."
inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
outputs = model(**inputs)
print(outputs.logits)

Model Performance

The model was evaluated using classification metrics such as accuracy, precision, recall, and F1-score. Detailed performance metrics can be found in the project documentation.

License

This model is open-source and available under the MIT License.

Acknowledgments

We acknowledge the creators of the OPPO-115 dataset for their contributions to privacy policy analysis research.

For any queries or collaborations, feel free to reach out to amlaninisarg@gmail.com or kavyaatulshahpatan@gmail.com.