links

Proposes modifications to the HITS algorithm that address link-spam vulnerabilities and topic drift by incorporating content similarity analysis and anchor text weighting into hub-authority score propagation. Experiments demonstrate that filtering semantically irrelevant links before iterative score computation reduces noise amplification, producing authority scores that more accurately reflect genuine topical relevance rather than raw link popularity. These refinements directly impact crawl prioritisation and authority-based ranking systems by making hub-authority scores resistant to manipulated link structures, improving the signal quality of link graph analysis for topical authority determination.