Effects of Prediction-Length on Accuracy in Automatic Assamese word prediction
Manash Pratim Bhuyan, Shikhar Kumar Sarma · 2019
Word prediction is a technique which increases the typing speed of user by suggesting the relevant word that the user is intended to write. In this research n-gram based language models are used to find the prediction accuracy for different size of prediction-list in Assamese language and the testing is done for both in-domain and mixed-domain text-data. Accuracy of prediction is calculated for a prediction-list length from 1 to 10 and the maximum accuracies obtained are 81.39% and 63.81% with prediction-list length 10 for in-domain and mixed-domain text-data respectively.