DeepURLBench Dataset

note README copied from source repo: https://github.com/deepinstinct-algo/DeepURLBench

This repository contains the dataset DeepURLBench, introduced in the paper "A New Dataset and Methodology for Malicious URL Classification" by Deep Instinct's research team.

Dataset Overview

The repository includes two parquet directories:

1. urls_with_dns:

   • Contains the following fields:
     • url: The URL being analyzed.
     • first_seen: The timestamp when the URL was first observed.
     • TTL (Time to Live): The time-to-live value of the DNS record.
     • label: Indicates whether the URL is malware, phishing or benign.
     • IP addresses: The associated IP addresses.

2. urls_without_dns:

   • Contains the following fields:
     • url: The URL being analyzed.
     • first_seen: The timestamp when the URL was first observed.
     • label: Indicates whether the URL is malware, phishing or benign.

Usage Instructions

To load the dataset using Python and Pandas, follow these steps:

import pandas as pd

# Replace 'directory' with the path to the parquet file or directory
df = pd.DataFrame.from_parquet("directory")

License

This dataset is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). You are free to use, share, and adapt the dataset for non-commercial purposes, with proper attribution.

Citation

@misc{schvartzman2024newdatasetmethodologymalicious,
      title={A New Dataset and Methodology for Malicious URL Classification}, 
      author={Ilan Schvartzman and Roei Sarussi and Maor Ashkenazi and Ido kringel and Yaniv Tocker and Tal Furman Shohet},
      year={2024},
      eprint={2501.00356},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2501.00356}, 
}