A PROBABILISTIC MODEL FOR SPELLING CORRECTION
Lucian Mircea Sasu · 2011
Spelling correctors are widely encountered in computer software applications and they provide the most plausible replacements for words that are presumably incorrect. The paper proposes a spelling correction method starting from the DamerauLevenshtein edit distance and using Bayesian decision theory. The resulted algorithm is tested on a bag of words from the New York Times news articles.