Open Access

ARTICLE

Twitter Arabic Sentiment Analysis to Detect Depression Using Machine Learning

Dhiaa A. Musleh, Taef A. Alkhales, Reem A. Almakki^*, Shahad E. Alnajim, Shaden K. Almarshad, Rana S. Alhasaniah, Sumayh S. Aljameel, Abdullah A. Almuqhim

1 Department of Computer Science, College of Computer Science and Information Technology, Imam Abdulrahman Bin Faisal University, P.O. Box 1982, Dammam, 31441, Saudi Arabia

* Corresponding Author: Reem A. Almakki. Email:

(This article belongs to the Special Issue: Machine Learning Empowered Secure Computing for Intelligent Systems)

Computers, Materials & Continua 2022, 71(2), 3463-3477. https://doi.org/10.32604/cmc.2022.022508

Received 10 August 2021; Accepted 11 October 2021; Issue published 07 December 2021

Abstract

Depression has been a major global concern for a long time, with the disease affecting aspects of many people's daily lives, such as their moods, eating habits, and social interactions. In Arabic culture, there is a lack of awareness regarding the importance of facing and curing mental health diseases. However, people all over the world, including Arab citizens, tend to express their feelings openly on social media, especially Twitter, as it is a platform designed to enable the expression of emotions through short texts, pictures, or videos. Users are inclined to treat their Twitter accounts as diaries because the platform affords them anonymity. Many published studies have detected the occurrence of depression among Twitter users on the basis of data on tweets posted in English, but research on Arabic tweets is lacking. The aim of the present work was to develop a model for analyzing Arabic users’ tweets and detecting depression among Arabic Twitter users. And expand the diversity of user tweets, by adding a new label (“neutral”) so the dataset include three classes (“depressed”, “non-depressed”, “neutral”). The model was created using machine learning classifiers and natural language processing techniques, such as Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), K-nearest Neighbors (KNN), AdaBoost, and Naïve Bayes (NB). The results showed that the RF classifier outperformed the others, registering an accuracy of 82.39%.

---

Keywords

---