A Study On The Detection And Generation Of Auto Correcting Suggestion For Real-Word Error In Mongolian Text

This paper is based on successful proofreading system of other languages to design and realized the text proofreading system with the functions of the checking of real-word error and the generation of correcting suggestion on the word-level error in Mongolian text.In the process of realizing the system,we made full use of the 10 million word-level "modern Mongolian corpus"created by the Mongolian College of Inner Mongolia University.We first extracted the word co-occurrence frequency from the corpus,and then developed the checking model of the real-word error on the basis of the word co-occurrence matrix.This paperresearches anddo an experiment from the following aspects.(1)There are mainly two modules of error checking and correction in the proofreading system of the Mongolian real-word error realized in this paper,error correction is the next step on the basis of error checking,so it is more important to study the checking of the Mongolian real-word error.Firstly,we used statistical methods to extract the co-occurrence frequency of the words from the 10 million word-level corpuses.Finally,we obtained 1,845,983 effective word frequency by statistics.(2)In order to provide the effective error correcting suggestions for the words that may be wrong,on the basis of the previous research results,we set up a dictionary of confusing words,which is called error correcting knowledge base.(3)On the basis of the word co-occurrence matrix,this paper uses the relationship of bi neighborship to design and realized the error checking algorithm of the Mongolian real-word error.Besides,this paper provides the correct suggestions for the Mongolian real-word error that has been found,and on the basis of the dictionary of confusing words,designs and realizes an algorithm of correcting suggestion-making.(4)On the basis of designing and implementing the system,the problems encountered in the whole process of the experiment and the corresponding solutions are summarized.And then this paper analyzes and evaluates the final experimental results.