COVID-19 Related Decision Making by Deep Learning Based Sentiment Analysis of Tweets

User sentiments shared in social media platforms like Facebook or Twitter significantly impacts the user's evaluation of the products or services. A novel technique for sentiment analysis using convolutional neural networks (CNNs) and bidirectional long short-term memory (BiLSTM) is proposed in this research. The proposed deep neural network model is trained using the COVID-19 pandemic Twitter dataset to separate a given tweet into three sentiment-based groups, namely (i) neutral, (ii) negative, and (iii) positive. Several experiments were conducted to evaluate the suggested model. The proposed model performs significantly in the results, with an accuracy of up to 97.895 present while comparing with the existing models.When compared against industry-leading baseline methods for tweet sentiment analysis. Insights into public mood and perception during the COVID-19 epidemic may be gained from this study, which shows the potential of sophisticated deep learning architectures to improve sentiment analysis accuracy.