Sentiment analysis on Education transformation during Covid-19 using Arabic tweets in KSA.

Abstract

The emergence of COVID-19 pandemic has changed the whole world. To prevent the spread of the virus, different precautionary measure and policies have been defined, one of them is distance learning. It has led to the educational transformation from physical education to online learning. Similarly, in KSA online education is adopted since March 2020. Inorder to extract the individual perception about the online education in KSA, twitter data was used. Arabic tweets were collected using twitter API. Furthermore, several machine learning models such as Support Vector Machine (SVM), Naïve Bayes (NB), Random Forest (RF), Logistic regression, K-nearest neighbors was used develop an automated sentiment analysis of online education related tweets in KSA. For feature extraction and selection N-gram and Term Frequency–Inverse Document Frequency (TF-IDF) was used. Logistic Regression achieved the highest accuracy of 69.33% for multiclass and Random Forest achieved accuracy of 80.35%. According to the dataset, most of the individuals have negative opinion about the online learning as the number of negative tweets are higher as compared to positive and neutral class