Improvement of the LR parsing table and its application to grammatical error correction

Abstract

A LR parsing table is generally made use of the parsing process based on the context free grammar for natural languages. Besides the parsing process, it can be used as the index of approximate pattern matching and error correction, because it has the characteristic to be able to predict the next character in the sentence. As for the issue of the traditional LR parsing table, however we can mention if the number of sequences to be processed becomes large, many reduce actions will be created in the parsing table, as a result, it takes a great deal of time to parse the sentence. In this paper, we propose the method to construct a new LR parsing table without reduce actions from the generalized context free grammar. This new parsing table denotes the states to be transited after accepting each symbol. Moreover, we applied this new parsing table to detect and correct erroneous sentences which include the syntax errors, unknown words and misspelling. By using this table, the symbol which is allocated just after the error position can be utilized for selecting correction symbols, as a result, the number of candidates produced on the correction process is reduced, and fast system can be realized. The experiment results, using 1050 sentences including error characters, show that this method can correct error points 69 times faster than the traditional method, also keep the almost same correction accuracy as the traditional method.