Monitoring People Emotions and Symptoms from Arabic Tweets during the COVID-19 Pandemic (Preprint)

Ali Al-Laith,Mamdouh Alenezi

DOI: https://doi.org/10.2196/preprints.24975

2020-10-13

Abstract:BACKGROUND COVID-19 started from Wuhan, China, in late December 2019, and is caused by the Corona Virus. It swept most of the world countries with confirmed cases and deaths. The World Health Organization (WHO) declared the virus as a pandemic on March 11th, 2020 due to its widespread transmission. A public health crisis was declared in specific regions and notional wide by governments all around the world. Citizens go through a wide range of emotions, such as fear of shortage of food, Anger at the performance of governments and health authorities in facing the virus, sadness over the death of a friend or relative, etc. OBJECTIVE We present a monitoring system of citizen’s concerns using emotion detection in Twitter data. We also track public emotions and link these emotions with COVID-19 symptoms. We aim to show the effect of emotion monitoring on improving people's daily health behavior and reduce the spread of negative emotions that affect the mental health of the citizens. METHODS We use Twitter API to collect and annotate 5.5M tweets in the period from January 2020 to August 2020. Two deep learning classifiers namely Convolutional Neural Network (CNN) and Long-short Term Memory (LSTM) employed to classify all tweets into six emotion classes (Anger, Disgust,Fear, Joy, Sadness, and Surprise) and two types (symptom and non-symptom tweets). RESULTS Our LSTM based text classification model outperforms the CNN model in emotion and symptom classification. We achieved a significant performance on multiclass classification (emotion detection) with an accuracy result of 91%. We also achieved an accuracy result of 88% on binary classification (symptom detection). The monitoring system shows that most of the tweets were posted in March. The anger and fear emotions have the highest number of tweets and user interactions after the joy emotion. The results of user interaction monitoring show that people use likes and replies to interact with non-symptom tweets while they use re-tweets to propagate tweets that mention any of the COVID-19 symptoms. CONCLUSIONS The study helps the governments and decision-makers to prove or deny these feelingsand discover other symptoms associated with the symptoms that were declared by the WHO. It can also help in the understanding of the people’s mental and emotional issues to address them before the impact of disease anxiety becomes harmful in itself.