Fuzzy matching

Use fuzzy matching when two strings may be nearly, rather than exactly, the same. IsFuzzyMatch compares a similarity score with a threshold between 0 and 1.

The typed overload makes the selected algorithm visible:

Matched := TStringKit.IsFuzzyMatch(
  'colour', 'color', 0.75, fmLevenshtein);

Method	Best starting point
`fmLevenshtein`	general spelling edits and insertions/deletions
`fmJaroWinkler`	short names where a common prefix matters
`fmLCS`	shared ordered subsequences

The legacy integer selector remains compatible (0 Levenshtein, 1 Jaro-Winkler, 2 LCS), but the TFuzzyMethod form is clearer in new code. The fuzzy match recipe is compiled with fmLevenshtein and a 0.75 threshold.

There is no universal right threshold. Collect representative matches and non-matches from your domain, then choose and test the threshold that gives the trade-off you need.