Publication

Abstract : Rapid development in digital data acquisition techniques have brought about massive extent of data. More than 80 percentage of these days records consists of unstructured or semi-structured data. The discovery of pattern and to analyze the textual content documents from massive volume of data is in trouble. Text mining is a prominent area of research as there is humongous amount of data already available. It is used to extract meaningful information from texts and has vast applications. Predicting the next word in a sentence, a text analysis application, is a challenging semantic modelling exercise. In this work, we have designed a model to automatically predict and suggest the next word to complete a sentence. To develop this model, we have used Latent Dirichlet Allocation, a topic modelling algorithm, for collecting predominant words under each topic and a Deterministic Finite Automata to maintain the syntactic meaning and structure of the sentence. The developed model is tested using a restaurant review dataset and validated in terms of coherence, perplexity, precision and recall which gives a competent prediction. © 2020 by Advance Scientific Research. This is an open-access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)