Multi-labelled Bengali Public Comments Sentiment Analysis with Bidirectional Recurrent Neural Networks (Bi-RNNs)
Promila Ghosh, Mohon Raihan, Nishat Tasnim Tonni, Himadri Sikder Badhon, Sayed Asaduzzaman, Hasin Rehana · 2023
A sentiment analysis is one of the most prominent research topics in today&s;s Natural Language Processing (NLP) field to analyze the statements or opinions of individuals. Individuals’ statements can be classified into different classes as positive, slightly negative, strongly negative or neutral. In this approach, multi-classified sentiments have been classified or analyzed using a Deep Learning (DL) algorithm named Bidirectional Recurrent Neural Networks (Bi-RNN) applied on Bengali text data. To analyze people&s;s comments on social sites or e-commerce sites sentiment analysis can play a notable role. The selected dataset for this approach has been multi-classified with mainly 7 types of sentiment tags based on the polarity to acknowledge the sentiment class with more specification. As one of the types of Recurrent Neural Networks (RNN) Bidirectional Recurrent Neural Networks (Bi-RNN) operates two RNN, the inputs are accessed in both forward and reverse directions. We have adopted multi-class classification with Bidirectional Recurrent Neural Networks (Bi-RNN) and acquired 88% exactitude.