Reciprocal Rank Fusion

Reciprocal rank fusion | Elasticsearch Guide 8.13 | Elastic $scor{e}_i=\frac{1}{a_i}+\frac{1}{b_i}$

• Same as harmonic mean for large and small relationships
• $k$ may be added
  • 
    • Reciprocal Rank Fusion outperforms Condorcet and individual Rank Learning Methods(PDF)
    • Here k=60
  • Increasing k reduces the difference between first and second place.

---

This page is auto-translated from [/nishio/Reciprocal Rank Fusion](https://scrapbox.io/nishio/Reciprocal Rank Fusion) using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.