Generation of Missing Words in Assamese text using N-gram based Model
Manash Pratim Bhuyan, Shikhar Kumar Sarma · Journal of Physics Conference Series · 2020
Abstract It is very common to miss certain words when writing while listening to others. A similar problem can arise when typing on the computer. The automatic generation of missed words shall very much helpful for users by suggesting the required words. In this research work, missed words of the Assamese sentences are generated, at present, there is no such tool/method exists which can provide or generate the missed words in an Assamese sentence. N-gram based models like bigram and trigram are used to generate missed words. Using the bigram and trigram models a rank is calculated for each possible suggested words and the suggested word list is sorted according to this rank value in decreasing order. Finally, these suggested words can be used to fill the place of the missed Assamese words. Different levels of experiments are carried out and the present proposed system can correct the missed words at an accuracy ranging from 58% to 66%. The proposed model can precisely generate accurate five relevant suggestions for a sentence containing an average of six words with two missed words separated by three words.