Sentiment Analysis of Arabic Tweets about Violence Against Women using Machine Learning
Moath Alzyout, Emran Al Bashabsheh, Hassan Najadat, Ahmad Alaiad
Abstract
Social Media platforms, such as Twitter became a significant pulse in smart societies that are shaping our communities by sensitizing people's information and perceptions across living areas over space and time. Social media sentiment analysis helps in recognizing people's emotions and attitudes and helps in assessing various public issues, such as, women's rights and violence against women. In this paper, we used the sentence based sentiment analysis to study the notion of women's rights. We collected Arabic dialect tweets from the whole Arab world as data via a Twitter API, then we cleaned the data to use it in the classification step. We have examined different types of traditional classification algorithms namely, Support Vector Machine, K-Nearest-Neighbour, Decision Trees, and Naive Bayes. Then, we compared these results with deep learning results. Finally, we compared the classification results using the precision, recall and accuracy measurements. We found that the Support Vector Machine algorithm gained the best results, while the Naive Bayes was the worst. We also noticed that there is an increasing attention to women's rights in the Arab world.