PARISlab at UCLA – Presentation, research interest, and publications – PARISlab at UCLA: Research, info, contact, and publications. · ranked by shared content topics & relevance
How the match score works
Each match is a 0–100 similarity score — the higher it is, the more two sites resemble one another.
It’s computed automatically from our own crawl data (never from what a site says about itself) by combining
several independent signals, so a high score means several of them point the same way:
- shared topics how many of the same fine‑grained content topics (IAB sub‑categories) both sites cover — the main signal.
- same category whether they sit in the same business / industry category.
- content & title how closely their titles and page descriptions line up in wording and meaning.
No single signal decides the result — they’re blended together. Treat the score as a way to rank
candidates rather than an absolute percentage; the chips on each result show which signals contributed.