puyang2025/phish_html

Phishing Dataset (Parquet)

Dataset Summary

This dataset contains HTML pages labeled as benign or malicious, collected from confirmed phishing URLs reported on PhishTank across 2018-2020. The data here is a direct conversion of the Kaggle dataset "phishing-dataset" by asifejazitu into a Hugging Face-compatible Parquet file.

Data Files

• train.parquet
• eval.parquet
• test.parquet

Data Fields

• text (string): HTML content of the page.
• label (class): benign or malicious.
• year (int32): year extracted from the folder structure (2018, 2019, 2020).
• path (string): original file path in the extracted dataset.

Data Splits

• train: 305,470 examples.
• eval: 38,184 examples.
• test: 38,184 examples.

Splits are created with stratified sampling to keep benign/malicious proportions consistent across splits.

Source Data

Original dataset on Kaggle: https://www.kaggle.com/datasets/asifejazitu/phishing-dataset

The authors note the pages were extracted from confirmed phishing URLs reported on PhishTank: https://phishtank.org/

Citation

If you use this dataset, please cite the original sources:

@misc{asifejazitu_phishing_dataset_2023,
  title = {Phishing Dataset},
  author = {asifejazitu},
  howpublished = {Kaggle Dataset},
  year = {2023},
  url = {https://www.kaggle.com/datasets/asifejazitu/phishing-dataset}
}

@misc{phishtank,
  title = {PhishTank},
  howpublished = {https://phishtank.org/}
}