A bigura-based real time sentiment analysis of new media

Abstract

Public opinion mining is an active research domain, especially the penetration of the internet and the adoption of smartphones lead to the enormous generation of data in new media. Thus generation of large amounts of data leads to the limitation of traditional machine learning techniques. Therefore, the obvious adoption of deep learning for the said data. A multilayer BiGura modal-based technique for real-time sentiment detection is proposed. The proposed system is analysed on different viral incidents such as Gaza’s invision. The exact case scenario is as follows “Taking Israel’s demand for millions of people from northern Gaza to migrate to the south”. In the experiment, the highest accuracy of the model in evaluating text content emotions and video content emotions reached 92.7% and 86.9%, respectively. Compared to Bayesian and K-nearest neighbour (KNN) classifiers, deep learning exhibits significant advantages in new media sentiment analysis. The classification accuracy has been improved by 3.88% and 4.33%, respectively. This research identified the fidelity of real-time emotion monitoring effectively capturing and understanding users’ emotional tendencies. It can also monitor changes in public opinion in real-time. This study provides new technical means for sentiment analysis and public opinion monitoring in new media. It helps to achieve more accurate and real-time monitoring of public opinion, which has important practical significance for social stability and public safety.