Search Relevance

finalScore =

α · score score + k + (1 − α) · Σi=1n wi · si

Normalized text relevance + normalized business signals.

01 / 06

How a result gets its score

Every hit is a convex combination of normalized text relevance and a weighted sum of business signals.

Projects

Explanation Search Debug Inspect how every term contributes to a result. Implementation Search Ranking Configure and calculate the transparent ranking formula.
Optimization Search Optimization Learn ranking parameters from weighted human relevance judgments. Under construction