The types of spelling errors in any language, including Korean, can be classified into the non-word and the context-sensitive spelling errors. Especially it should be considered that the semantic and the syntactic relationships in the context to find ... The types of spelling errors in any language, including Korean, can be classified into the non-word and the context-sensitive spelling errors. Especially it should be considered that the semantic and the syntactic relationships in the context to find the context-sensitive spelling errors. So, finding and correcting context-sensitive spelling errors has a higher level of difficulty than that of the non-word spelling errors. An earlier statistical method used in correcting statistical context-sensitive spelling errors is based on the word frequency. However this method has vulnerability that it may cause difficulty in distinguishing words of similar meaning. This proposes a way to solve the problem of this method. The proposed method adopts the vocabulary expanding scheme using relational words of Korean Word. This method also resolves data sparseness problems and can effectively distinguish words of similar meaning. The experiment on Sejong corpus shows that the proposed method is more effective than Laplace smoothing in that the precision and the recall have been measured to be 1.46% and 3.93% higher than those of Laplace smoothing, respectively. ，韩语论文网站，韩语论文 |
