PDNS-Net
收藏资源简介:
PDNS-Net是由斯里兰卡科伦坡大学和卡塔尔计算研究机构共同创建的大型异构图数据集,包含447,000个节点和897,000条边,用于恶意域名分类任务。该数据集通过被动DNS数据收集构建,涵盖了2021年10月的互联网域名解析信息。PDNS-Net不仅规模远超同类数据集,还提供了详细的数据收集方法、异构图构建、描述性统计和初步的图分类性能分析。该数据集的应用领域主要集中在网络安全,特别是用于改进图学习算法在大型异构图上的性能,解决恶意域名识别等问题。
PDNS-Net is a large-scale heterogeneous graph dataset jointly created by the University of Colombo in Sri Lanka and the Qatar Computing Research Institute. It contains 447,000 nodes and 897,000 edges, and is designed for malicious domain classification tasks. This dataset is constructed via passive DNS data collection, covering Internet domain name resolution information from October 2021. PDNS-Net not only far exceeds similar datasets in scale, but also provides detailed data collection methods, heterogeneous graph construction procedures, descriptive statistics, and preliminary graph classification performance analyses. Its application fields mainly focus on cybersecurity, specifically for improving the performance of graph learning algorithms on large-scale heterogeneous graphs and solving problems such as malicious domain name identification.




