Detection and correction of real-word spelling errors in Persian language

Several statistical methods have already been proposed to detect and correct the real-word errors of a context. However, to the best of our knowledge, none of them has been applied on Persian language yet. In this paper, a statistical method based on mutual information of Persian words to deal with context sensitive spelling errors is presented. Different experiments show the accuracy of correction method on a test data which only contains one real-word error in each sentence to be about 80.5% and 87% with respect to precision and recall metrics.