DataTerminology.comNEW
Also known as: Hyperlink-Induced Topic Search
A link analysis algorithm that computes hub and authority scores for nodes from the structure of hyperlinks.
DESCRIPTION
HITS iteratively assigns two scores to each page: authority (valuable content) and hub (links to authorities). It leverages the adjacency structure of the web, performing alternating updates that converge to principal eigenvectors. Developed by Jon Kleinberg, HITS influenced early web search ranking and analysis of citation networks. It is sensitive to topic drift and often complemented by PageRank for robustness.
ALTERNATIVE DEFINITIONS
Wikipedia
Hyperlink-Induced Topic Search (HITS) is a link analysis algorithm that rates Web pages.
Defines HITS succinctly and emphasizes its role in rating pages.
Jon Kleinberg
We develop a method for identifying authoritative pages, by analyzing the link structure of the Web.
Original research framing authority and hub concepts via link analysis.
Stanford IR Book
HITS assigns to each node two scores, authority and hub, computed from the link structure.
Educational text focusing on the dual scoring mechanism.
RELATED TERMS
Link Analysis Graph Algorithms Authority Score Hub Score Eigenvector Centrality Web Search Adjacency Matrix Spectral Methods
COMPETING TERMS
PageRank SALSA
KNOWLEDGE AREAS
AI Data Science Mathematics Semantic & Knowledge Modeling
TARGET AUDIENCE
Data Scientist AI Engineer Software Developer
EXTERNAL SOURCES
- Wikipedia - HITS algorithm
- Jon Kleinberg - Original HITS paper
- Stanford IR Book - Chapter on link analysis
- NetworkX - HITS algorithm implementation docs
TOP QUESTIONS
Who invented the HITS algorithm?
Jon Kleinberg introduced HITS, framing authority and hub scores derived from hyperlink structure.
Category: Definition
How does HITS differ from PageRank?
HITS computes dual scores (hub and authority) via local subgraph analysis; PageRank computes a global popularity score via random walks.
Category: Alternatives
What is topic drift in HITS?
HITS can converge to off-topic authorities due to subgraph selection; careful seed selection mitigates drift.
Category: Risks
Where is HITS used today?
Citation networks, recommendation, and link-based analytics in constrained domains benefit from HITS’s dual-role insights.
Category: Example
What are common mistakes implementing HITS?
Ignoring normalization and convergence criteria can yield unstable scores; use robust iteration and damping if needed.
Category: Mistakes
How does HITS compute hub and authority scores?
Authorities are proportional to inbound hub scores; hubs to outbound authority scores; iterative updates reach fixed points.
Category: Definition
Is HITS sensitive to link spam?
Yes; adversarial linking can inflate hubs or authorities. Combine with trust propagation or spam filters.
Category: Risk
Can HITS be applied to non-web graphs?
Absolutely, any directed graph with notion of endorsements can leverage HITS scores.
Category: Example
What is SALSA compared to HITS?
SALSA blends HITS and random walk ideas, often more robust to topic drift.
Category: Alternatives