Detecting COVID-19 Implication on Education and Economic in Arab World Using Sentiment Analysis Techniques of Twitter Data
Omar AlZoubi, Farah Shatnawi, Saif Rawashdeh, Muneer Bani Yassein, Ismail Hmeidi · 2022
Coronavirus (COVID-19) is a contagious and dangerous infection that initially surfaced in Wuhan, China, in December of 2019 and has infected millions around the world. The rapid transmission of this disease leads to all governments’ taking a lockdown decision. This decision has a negative effect on the two main sectors of any country: education and the economy. This study explores public opinion about the impact of COVID-19 on these two sectors by collecting Arabic tweets from Twitter from the MENA region. Then, we applied NLP and sentiment analysis process to the dataset for each sector after applying the pre-processing steps. We built state-of-the-art machine learning models to use them in the prediction process that are Random Forest (RF), Logistic Regression (LR), Na¨ıve Bayes (NB), and Voting classifier (VC). The results showed that the NB and VC gave the best performance results in two datasets. The results showed that these models produced outcomes that were useful to many decision-makers in the government in taking important and suitable decisions related to daily life in the Arab World.