DeepInstinct/DeepURLBench

DeepURLBench是一个大规模的现实世界URL分类基准数据集，由Deep Instinct研究团队开发。数据集包含两个子集：urls_with_dns和urls_without_dns，分别包含不同的元数据级别。该数据集旨在用于网络安全和机器学习的研究和教育目的，并遵循CC BY-NC 4.0许可。数据集没有预定义的训练/验证/测试分割，用户应按first_seen字段进行时间分割。该数据集应仅以只读方式使用，并且强烈警告不要与数据集中的任何URL进行交互。

DeepURLBench is a large-scale benchmark dataset for real-world URL classification, developed by Deep Instincts research team. The dataset includes two subsets: urls_with_dns and urls_without_dns, each with different levels of metadata. It is intended for research and educational purposes in cybersecurity and machine learning, and is licensed under CC BY-NC 4.0. The dataset does not include predefined train/validation/test splits and should be split chronologically by the first_seen field. The dataset is meant to be used in a read-only context, and there is a strong warning against interacting with any URLs in the dataset.