Automatic Detection of Coronaphobia in Twitter: Study Case of Republic of Panama

Social networks spread their content quickly, and the online opinions of users are valuable for decision-making. The quarantine imposed by the Covid-19 pandemic has had a negative impact on the physical and mental health of the Panamanian population. Anxiety, fear and stress are common psychological reactions among the population today, due to the uncertainty experienced; these reactions or phobias experienced by the population are understandable and are called Coronaphobia. During the quarantine, social networks such as Twitter were the means of communication for millions of people. The purpose of social networks was to share opinions or comments, making it possible to have and analyse a large amount of information. Analysing tweets requires a systematic process of collecting, transforming and classifying them. This requires the use of artificial intelligence tools and other innovative techniques. Sentiment analysis is the task of determining the emotional tone behind a sentence in order to extract meaningful information from users related to their opinions, emotions and attitudes. Therefore, our study has used a Spanish corpus extracted from Twitter to analyse the sentiment of opinions and comments in tweets and to find patterns that indicate Coronaphobia conditions in the Panamanian population, using artificial intelligence techniques. According to the results obtained, we can say that the final model provides competitive results in terms of automatic sentiment identification, with an accuracy of over 93%.