Fuzzy-name-matching-and-conversion

A fuzzy matching toolkit for noisy names and transliterations. Includes normalization, phonetic encoding, candidate generation, and ranking to improve recall and precision when matching names with spelling variants and script conversions.

Highlights

• Name normalization and cleanup utilities
• Phonetic keys for sound-alike matching
• Edit-distance scoring and ranking
• Candidate generation to keep search efficient

Approach

1. Normalize input strings (case, spacing, punctuation, common variants)
2. Generate phonetic keys to capture sound-alike matches
3. Create candidates using prefixes, phonetic buckets, or blocking
4. Rank results using edit distance and weighted scoring

Project Structure

• src/ - core matching and conversion logic
• examples/ - sample inputs and usage
• tests/ - unit tests (optional)
• data/ - sample datasets (optional)

Usage

• Provide two names (or a query and a candidate list)
• The library returns top matches with scores and intermediate signals (optional)

Performance Notes

• Candidate blocking is used to reduce comparisons
• Scoring can be tuned based on the domain (names, locations, entities)

License

Add your preferred license.