-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Description
Goal
Implement BM25 (Okapi) ranking algorithm.
Learn
- How BM25 improves on TF-IDF
- Document length normalization
- k1 and b parameters and what they control
- When BM25 beats TF-IDF
Deliverable
- Working BM25 ranker with parameter tuning
- Comparison with TF-IDF on same queries
- Analysis of parameter effects
Proof Point
Can explain the formula and tune parameters for different use cases.
Directory
search-fundamentals/02-bm25-implementation/
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels