Context-associative approach to automated correction of spelling errors

Abstract

Addition of filtration for versions of errors correction (by the semantic criterion) into the spelling correction scheme provides a high level of programmatic proof-reader accuracy and makes its functioning possible. The deviation from the traditional sequence of text data analysis allows one to perform a context-associative processing the surroundings of a distorted word at any stage of spelling correction. Techniques for evaluation of the correction algorithm accuracy are offered and up-to-date problems concerning realization of the context-oriented spelling correction are determined. They will be studied in further work.