Cricket Sentiment Analysis from Bangla Text Using Recurrent Neural Network with Long Short Term Memory Model

Abstract

Nowadays, people used to express their feelings, thoughts, suggestions and opinions on different social platform and video sharing media. Many discussions are made on Twitter, Facebook and many respective forums on sports especially cricket and football. The opinion may express criticism in different manner, notation that may comprise different polarity like positive, negative or neutral and it is a challenging task even for human to understand the sentiment of each opinion as well as time consuming. This problem can be solved by analyzing sentiment in respective comments through natural language processing (NLP). Along with the success of many deep learning domains, Recurrent Neural Network (RNN) with Long-Short-Term-Memory (LSTM) is popularly used in NLP task like sentiment analysis. We have prepared a dataset about cricket comment in Bangla text of real people sentiments in three categories i.e. positive, negative and neutral and processed it by removing unnecessary words from the dataset. Then we have used word embedding method for vectorization of each word and for long term dependencies we used LSTM. The accuracy of this approach has given 95% that beyond the accuracy of previous all method.