Implement "fuzzy" name matching - Githubissues

Logiqx / sse-results

Speed Sailing Event Results

https://logiqx.github.io/sse-results/

GNU General Public License v3.0

0 stars 0 forks source link

Implement "fuzzy" name matching #23

Closed Logiqx closed 2 years ago

Logiqx commented 2 years ago

Implement fuzzy matching to detect typos or variations of names across different years.

Calculate the Levenshtein distance using the full name(s), deeming a distance < 5 as a potential match.

Example typos to detect:

Graham Holbert vs Graham Hulbert (distance 1)
Trevor Watford vs Trevor Whatford (distance 1)
Mathew Spooner vs Matthew Spooner (distance 1)
Kin Newman vs Kim Newman (distance 1)
Piere Saville vs Pierre Saville (distance 1)
Jon Montgomery vs John Montgomery (distance 1)
Alistair Williams vs Alastair Williams (distance 1)

Example short names to detect:

Pete Martin vs Peter Martin (distance 1)
Dave MacInnes vs David MacInnes (distance 2)
Dan Robinson vs Daniel Robinson (distance 3)
Jim Paine vs James Paine (distance 3)
Matt Spooner vs Matthew Spooner (distance 3)
Eddie Murrell vs Edward Murrell (distance 3)
Ed Murrell vs Edward Murrell (distance 4)
Robert Date vs Bob Date (distance 4)

Example variations to detect:

Hans-Juergen vs Hans-Jürgen (distance 2)
Claude van Martyn vs Claude Van-Martyn (distance 2)
Claude Van-Martyn vs Claude Van-Man-Martyn (distance 4)

Logiqx commented 2 years ago

Implemented this on 17/12/2021 and 18/12/2021