ThreatCrawl
收藏资源简介:
本文提出了一种名为ThreatCrawl的系统,旨在通过多臂老虎机(MAB)和多种爬取策略,动态地识别和爬取与网络威胁情报(CTI)相关的网页。该系统使用SBERT嵌入技术来识别相关文档,并根据内容动态调整爬取路径。ThreatCrawl系统在爬取过程中实现分类,并且能够维持主题焦点,有效地扩展了种子集合。该数据集主要用于评估ThreatCrawl系统的有效性,其中包含了多种爬取策略组合的结果,如前向链接搜索、后向链接搜索和关键词搜索等。数据集的大小、数据量和Tokens数等信息在论文中并未提及。
This paper proposes a system named ThreatCrawl, which aims to dynamically identify and crawl webpages related to Cyber Threat Intelligence (CTI) via Multi-Armed Bandit (MAB) and various crawling strategies. This system employs SBERT embedding technology to identify relevant documents and dynamically adjust crawling paths based on content. The ThreatCrawl system performs classification during the crawling process, maintains thematic focus, and effectively expands the seed set. This dataset is primarily used to evaluate the effectiveness of the ThreatCrawl system, and contains results from multiple crawling strategy combinations, such as forward link search, backward link search, and keyword search, among others. Details such as the dataset size, data volume, and number of Tokens are not mentioned in this paper.




