covid-19 pandemic sentiment analysis using deep learning methods

In the last two decades, humans have been exposed to viruses and pandemics, each of which has had a profound effect on the life and activities of individuals, but today the coronavirus has the greatest impact on the social behavior of different societies, a significant part of which can be viewed on social media platforms like twitter. however, in such a critical situation, twitter has become a platform where people express their feelings and opinions about covid-19 pandemics, and a sentiment analysis of these texts can be understood how people react to a virus. in this study, sentiment analysis as a deep learning technique is used to detect the polarity within text data. for this purpose, this paper uses tweets covid sentiment values which are related to the covid-19 pandemic and contain the text of tweets and their polarity values. furthermore, this study conducts a series of experiments with long short-term memory (lstm) as one of the types of recurrent neural network (rnn) architectures. the results confirm that the lstm model with designed embedding layer has better result with mae and mse. the results indicate why some tweets have positive polarity and other negative polarity.