Measuring the evolution of social hope since the outbreak of COVID 19 in Iran with a case study of Persian Twitter based on computational social science

Social hope is a category that has received some attention in recent years, and theoretical constructs have yet to be developed to measure it. In this sense, one of our goals in this research is a measurable definition of social hope. On the other hand, big data has recently been introduced to the social sciences and humanities, which has attracted the attention of researchers regarding methods, techniques, and the required tools for working with it. In this research, we try to use the big data available in the Twitter social network to measure social hope after reviewing theories in the category of social hope. In this regard, the tweets about the Covid-19 virus have been collected since its outbreak in Iran and analyzed based on the existing and emerging theoretical principles. The analysis of more than one million tweets shows that social hope decreases as this virus spreads in the country. At the same time, our time slices analysis show that people’s attitude toward their past varies at different times, and people may reconstruct their social past concerning the current situation. An approach based on the Capsule network and XGBoost model is used to classify these tweets, and this model is able to achieve an accuracy of 0.9640 in Polarity detection and 0.9997 in Domain detection.