Ensemble learning and stacked convolutional neural network for Covid-19 situational information analysis using social media data

Twitter has evolved into a primary platform for sharing opinions and insights on current events, such as the ongoing coronavirus pandemic. It’s now regarded as one of the foremost sources for conducting analyses, making predictions, and extracting valuable knowledge for machine learning endeavors. it’s crucial for the public and authorities to have access to timely and relevant situational information. This enables them to respond effectively and make informed decisions. This paper categorizes situational information into seven distinct groups and examining the behavioral and emotional shifts that individuals are experiencing amid the Covid-19 pandemic. In this paper, we have done analysis of real-time sentiment in tweets related to the Covid-19 pandemic. We have collected the historical Covid-19 tweets from 01/10/2020 to 30/03/2021. We have proposed an adaptive ensemble learning and Stacked-CNN model and applied the proposed models on the collected datasets for Covid-19 situational information analysis. The experimental results demonstrate the effectiveness of the Adaptive ensemble model and the Stacked CNN model in providing better predictions for real-time sentiment analysis of tweets related to the Covid-19 pandemic. The proposed methods can provide valuable insights into public opinion and emotions and has practical applications in areas like public health communication, crisis management, and understanding community concerns.