DeepURLBench
收藏资源简介:
DeepURLBench是由Deep Instinct和本古里安大学联合创建的一个多类恶意URL分类数据集,旨在提升网络安全领域的URL分类模型性能。该数据集包含超过2200万条URL,分为良性、钓鱼和恶意三类,数据来源于2020年至2023年间的公开拒绝列表和允许列表,并通过VirusTotal进行标签标注。数据集经过严格的清洗和结构化处理,确保了数据的多样性和时效性。该数据集的应用领域主要集中在网络安全,旨在通过深度学习和传统机器学习方法提升恶意URL的实时分类能力,解决现有模型在实时性和准确性上的不足。
DeepURLBench is a multi-class malicious URL classification dataset jointly created by Deep Instinct and Ben-Gurion University, aiming to improve the performance of URL classification models in the field of cybersecurity. This dataset contains over 22 million URL entries, categorized into three classes: benign, phishing, and malicious. The data is sourced from public allowlists and blocklists spanning the period from 2020 to 2023, with labels annotated via VirusTotal. The dataset has undergone rigorous cleaning and structuring processes to ensure its diversity and timeliness. Its primary application scenarios focus on cybersecurity, aiming to enhance real-time classification capabilities for malicious URLs through deep learning and traditional machine learning methods, and address the shortcomings of existing models in terms of real-time performance and accuracy.

- 1A New Dataset and Methodology for Malicious URL ClassificationDeep Instinct 和 本古里安大学 · 2024年



