Seonglae ChoHyperlink-Induced Topic Search (HITS) Model
Unlike PageRank, HITS is calculated on a per-query basis and demonstrates superior search performance for specific topics. It trying to solve Abundance Problem by selecting pages based on authority and hub. The model identifies two types of important pages in a hyperlinked environment:
Authority Pages with Authority score
A good authority has links from many good hubs
- Pages that receive links from many high-quality hub pages
- Considered authoritative sources for specific topics
Hub Pages with Hub score
A good hub links to many good authorities
- Pages that link to many high-quality authority pages
- Act as curated lists or directories of authoritative content
This will turn this into an Circular definition architecture will turn this into an iterative computation
1. Root Set
Given text query, use a text index to get all pages containing query.
2. Base Set
Add in any page that either points to a page in the root set, or is pointed to by a page in the root set.
3. HITS Iteration
just as Multi-stage Recommender
HITS on the Web: How does it Compare? - Microsoft Research
This paper describes a large-scale evaluation of the effectiveness of HITS in comparison with other link-based ranking algorithms, when used in combination with a state-of-the-art text retrieval algorithm exploiting anchor text. We quantified their effectiveness using three common performance measures: the mean reciprocal rank, the mean average precision, and the normalized discounted cumulative gain measurements. […]
https://www.microsoft.com/en-us/research/publication/hits-on-the-web-how-does-it-compare/