Analyzing Social Media Opinions Using Data Analytics
This project collected Arabic tweets about a certain event that happened in 2018 and showed that the Voting classifier with SVM in Python gives the best performance, in terms of accuracy, precision, and recall, as compared with the other algorithms.
Abstract
In recent years, Twitter has become one of the most prevalent social media websites, in which people in the Arab world try to express their feelings and opinions about several political, social and economic events in their daily lives. In this project, we collected Arabic tweets about a certain event that happened in 2018. We preprocessed and analyzed the data to extract the key information. We used several algorithms to construct the classification model, including Naïve Bayes (NB), Support Vector Machine (SVM), Decision Tree (DT), K-Nearest Neighbour (K-NN) and Ensemble Classifiers (ex: Bagging, Voting) that combined these algorithms. The results showed that the Voting classifier with SVM in Python gives the best performance, in terms of accuracy, precision, and recall, as compared with the other algorithms.