Phishing Website URLs with Binary Classification Labels

A subset of phishing site data originally sourced from Kaggle, used for a model compression example. The dataset was uploaded by user 'shawhin' to Hugging Face and last updated on September 1, 2024. It contains website URLs labeled as phishing (1) or not phishing (0).

Use Cases

• Training binary classifiers to detect phishing websites based on URL text.
• Benchmarking model compression techniques like knowledge distillation and quantization.
• Developing lightweight security models for real-time URL scanning.

Strengths

• Provides a clear binary label (phishing or not) for each URL.
• Includes three associated BERT-based models demonstrating a model compression workflow.
• Has a documented origin, being a subset of a Kaggle dataset.

Limitations

• Row count, column details, and file formats are unknown, limiting suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Subset of data from Kaggle dataset 'Phishing Site URLs' by taruntiwarihp.

Collection Method

Likely collected from web sources for cybersecurity research.

Time Range

null

Freshness

Last updated 2024-09-01 20:03:03; freshness should be verified.

Geography

null