Error Recovery in Derivation of Context-Free Grammars Applied to Correction of Spelling and Grammatical Errors
Natural Language Processing, POS tagging, Error Correction, Top-Down Parsing, Lexical Analysis
Natural Language Processing (NLP) is an interdisciplinary field that studies computational processing of natural speech. It is a growing area due to the popularity of Chat and Voice Interfaces. NLP application are built upon a basic flow of steps: tokenization, stemming, part-of-speech (POS) tagging, stop-word removal, dependency analysis, named entity recognition and co-referencing, and while those steps already work really well on English texts, the same cannot be said for every language. This paper investigates the use of techniques developed to correct errors in compilers to improve the result of POS-tagging in phrases with classification problems. The results presented in the research indicate that the developed process can be used effectively in some contexts.