Fankestein's insects (abstract)

Research aimed at correcting words in text has focussed on threeprogressively more difficult problems: 1) non-word error detection;2) isolated-word error correction; and3) contextdependent word correction.In response to the first problem, efficient pattern matching andn-gram analysis techniques have been developed for detectingstrings that do not appear in a given word list. In response to thesecond problem, a variety of general and application-specificspelling correction techniques have been developed. Some of themwere based on detailed studies of spelling error patterns. Inresponse to the third problem, a few experiments using naturallanguage processing tools or statis tical language models have beencarried out. This article surveys documented findings on spellingerror correction techniques, reviews the state of the art ofcontext-dependent word correction techniques, and discussesresearch issues related to all three areas of automatic errorcorrection in text. (The Article on which the Talk was based appeared in the December1992 Issue of Computing Surveys) 515