Sentiment Analysis of Social Media's Bangla Data

Abstract

Sentiment analysis is a way where huge amount of data can be categorized into different sentiments, emotions, attitudes or opinions. Using sentiment analysis sentiments or emotions can be predicted from individual’s viewpoint which means individual, and public are expressing their views and from using these views their sentiments can be detected using sentiment analysis. Recently sentiment analysis has been performed on various information of social media to derive market intelligence. As we know social media is filled with different contents and audiences are interacting there making it a huge opportunity to perform sentiment analysis on this information. In terms of Bengali contents, many audiences there interact with Bengali language which makes it a treasure trove to perform sentiment analysis in Bengali NLP field. Here in this study sentiment analysis has performed on the audience’s Bengali comments expressing different views of regarding social media’s Bengali contents. The dataset containing these public views are 4000 Bengali comments collected from Facebook and YouTube Bengali contents. Here positive, negative and neutral classes are used to categorize the Bengali data and tokenizer from Keras library is used to tokenize the Bengali text. Deep learning is part of that machine learning algorithm where its structures are inspired by the human brain. In this study deep learning algorithm LSTM and Bi-LSTM are used where Keras library is used to run these algorithms. Deep learning algorithm LSTM and Bi-LSTM are performed, and Bi-Directional LSTM has the highest accuracy of 97.25% than LSTM.