State-of-the-art spelling correction algorithms

Question

The most successful algorithm to date is Andrew Golding and Dan Roth's "Winnow-based spelling correction algorithm", published in 1999, which is able to recognize about 96% of context-sensitive spelling errors, in addition to ordinary non-word spelling errors.

Is this up-to-date? What if we limit to algorithms not covered by patents?

EDIT: After reading the papers on Winnow, it only seems to support distinguishing a limited number of confusion sets, which need to be chosen before training, and so can't serve for general spelling correction anyway.

score 3 · Answer 1 · answered Jul 16 '13 at 12:59

From an applicative point of view the following software solutions are quite popular:

GNU Aspell (which replaces Ispell)
Hunspell, look at the quite impressive list of uses, including Apache Solr, Google Chrome, Firefox, etc.

Finally, Enchant is an interesting aggregator as it capable of having multiple backends loaded at once.

State-of-the-art spelling correction algorithms

1 Answers1