A Text Based Sentiment Analysis Model using Bi-directional LSTM Networks

Abstract

The Web generates enormous quantities of data in the form of user attitudes, feelings, thoughts, and opinions about various social occasions, brands, manufacturers, and democracy through social media platforms, communities, online reviews, and blog posts. Users' opinions expressed on the internet have a significant impact on users, product manufacturers and legislators. Sentiment analysis is stated as text association which categorizes the conveyed ethos or state of mind in variousbehaviours such as good, bad, satisfactory, hostile, like, dislike, etc. Mostly in domain of Speech Recognition, the lack of sufficient marked information is critical with sentiment classification. To overcome this problem, sentiment analysis is performed using RNN as deep NN are successful due to its machine learning abilities. This paper highlights a methodology for sentiment classification by means of a deep NN model called BiLSTM recurrent neural networks for a variety of sentiment analysis challenges, including text categorization, cross-lingual issues, verbal and contextual assessment, and industry news assessment, among others.