MURL: A Large-Scale Multi-Class Dataset for Malicious URL Detection and Cyber Threat Classification

This dataset contains 783,705 URLs categorized into eight classes: Benign, Phishing, Defacement, Malware, Legitimate, Spam, and Exploit. The dataset is designed to support research in malicious URL detection, cybersecurity, cyber threat classification, machine learning, and deep learning. Its diverse class distribution also makes it useful for studying multi-class classification and class imbalance in cybersecurity applications.本数据集包含783,705个统一资源定位符（URL），分为八类：良性（Benign）、网络钓鱼（Phishing）、网页篡改（Defacement）、恶意软件（Malware）、合法（Legitimate）、垃圾链接（Spam）与漏洞利用（Exploit）。该数据集专为支撑恶意URL检测、网络安全、网络威胁分类、机器学习及深度学习领域的研究而设计，其多样化的类别分布还可用于研究网络安全应用场景下的多分类任务与类别不平衡问题。