A Naive Bayes Approach for Word Prediction using NLP

Word prediction is a very important natural language processing downside during which we would like to predict the proper word in an exceedingly given context. Word completion utilities, prophetical text entry systems, writing aids, and language translation square measure a number of common word prediction applications. This paper presents a brand-new word prediction approach supported context options and machine learning. The planned technique casts the matter as a learning-classification task by coaching word predictors with extremely discriminating options designated by numerous feature choice techniques. The contribution of this work lies within the new approach of presenting this downside, and also the distinctive combination of a prime performing artist in machine learning, svm, with numerous feature choice techniques MI, X2, and more. The tactic is enforced and evaluated victimization many datasets. The experimental results show clearly that the tactic is effective in predicting the proper words by utilizing little contexts. The system achieved spectacular results, compared with similar work; the accuracy in some experiments approaches ninety-one correct predictions.