Next Word Prediction Using Deep Learning: A Comparative Study
Milind Soam, Sanjeev Thakur
Abstract
Deep learning is a subclass of machine learning, it mimics the functionality of the humanbrain the way it processes and creates pattern in the facts for choice making. It is basically an AI function that has networks capable of learning unsupervised data that is shapeless. The succeeding word forecast is performed on dataset consisting of texts. Next Word Prediction is an application of NLP (Natural LanguageProcessing). It is also known as Language Modelling. Basically it is the process of predicting the next word in a sentence. It has many applications which are usedby most of us such as auto-correct which is mostly usedin Emails / Messages; it has also it’s usage in MS Wordor google search where forecasts the next word based on our search history or the search we did for globe. In this work we have studied NLP, different deep learning techniques such as LSTM, BiLSTM and performed a comparative study. We found satisfactory results in BiLSTM and LSTM [1]The accuracy received using BiLSTM and LSTM are: 66.1% and 58.27% respectively.