A Smart Proofreader for All Natural Languages: Achieving Semantic Understanding by Majority Vote. (19th February 2012)

• Record Type: Journal Article
• Authors: Olsen, Kai A.
• Other Names: Gentile C. Academic Editor.
• Is Part Of: ISRN artificial intelligence. Volume 2012(2012)
• Publication Date: 2012-02-19
• Subjects: Artificial intelligence -- Periodicals; Artificial intelligence; Periodicals; 006.3
• DOI: 10.5402/2012/918362
• Languages: English
• ISSNs: 2090-7435

Abstract

Abstract : The language tools offered in common word processors use dictionaries and simple grammatical rules. They cannot detect errors such as a wrong preposition, interchanged words, or typos that result in a dictionary word. However, by comparing the user's text to a large repository, it is possible to detect many of these errors and also to suggest alternatives. By looking at full sentences, it is often possible to get the correct context. This is important in detecting errors and in order to offer valuable suggestions. These ideas have been implemented in a prototype system. We present examples in English and Norwegian, but the method, that of following a "majority vote, " can be applied to any written language.