Deep Learning for Discourse Analysis-Based Sentiment Analysis of Tweets

Through innovative days, too much information is created, and a massive quantity of information in the network is meant for internet users through innovative network tools with development. Creativity is replaced with opinion sharing for online education, and the internet has become an essential stage. People can exchange and deliver their perspectives on subjects and conversations with various groups; otherwise, they can post and send information around the globe with public internet spots such as Twitter, Facebook. Here, we have explained discourse analysis concerning understanding the language and extracting the hidden meaning in Twitter sentences. Also, sentiment analysis is an NLP technique for knowing the opinions of consumers. Twitter information is used to predict tweet data, whether the opinions are highly unstructured, precisely as good, bad, or impartial. Initially, preprocessing is done, and then the classification technique Convolution Neural Network (CNN) is used to classify preprocessed tweet datasets as well as extract the people’s emotions and predict the comments like positive, negative, and neutral with training 80% and testing 20% results are compared with proposed and existing techniques. The proposed CNN gives 94.23% accuracy with better results than existing techniques.