AS2Biz: Leveraging Web Presence and AI to Improve AS Business Classification

Abstract

Classifying Autonomous Systems (ASes) by the business sectors of their operating organizations, also referred to as AS business classification, is a relatively recent but increasingly important area of study. However, the only publicly available dataset for this purpose suffers from significant inaccuracies and limited maintainability due to its heavy reliance on third-party sources. We argue that an organization's web presence—particularly its official website—serves as a more reliable and transparent source for business classification, offering broad coverage and regularly updated information. We introduce AS2Biz, a web presence-based methodology for AS business classification. AS2Biz first constructs an AS-to-website mapping dataset, achieving 98.5% coverage and 94.4% accuracy, and then mainly leverages large language models to classify textual website content into business categories. The entire methodology is highly automated and cost-efficient. Our final AS2Biz dataset covers 94.3% of ASes and, on a 220-AS ground truth benchmark, achieves substantially higher accuracy—more than doubling the F1 score on subdivided business categories (0.78 vs 0.37) and improving the F1 score on general business sectors (0.87 vs 0.62).