Word correction using noisy channel

Abstract

In now days we are continuously using social media. Due to the massive uses of social media, we are typed messages different posts in social media. Due to the massive typing sometimes we are write misspelled words in the place of any correct word. The Aim of this thesis is to detect the misspelled word and the error type. After that the word corrected by noisy channel model and suggest some probable best word against of this misspelled word. For the correction we uses noisy channel model and one large corpus for the correction. After that we detect the misspelled word between the sentence and correct this word through noisy channel model and suggest the bigram probability of this sentence. In this project I will describe the process of this correction.