Prediction of depressed Arab women using their tweets

ABSTRACT

Social media is increasingly used to share thoughts and feelings. By mining the social media posts, a picture of users’ behavior may be obtained, to predict mental illnesses, like depression. Depression is seen as a taboo in the Arab world and cannot be discussed publicly out of fear of criticism. This led many to express their feelings over social media as an escape. Research has been conducted on the ability to use machine learning to make an early-stage diagnosis of depression. However, there is not enough research on this in the Arab world. This research was conducted on Arab women’s tweets during the COVID-19 pandemic to predict whether they have depression symptoms using machine learning. The proposed contribution is to create a Recurrent Neural Network (RNN) model to predict depression from tweets. The model is evaluated using 10000 tweets extracted from 200 users and the obtained results shows its effectiveness.

KEYWORDS:

• Machine learning
• depression
• prediction
• Arabic language
• Twitter
• public health

Acknowledgment

This research was funded by the Deanship of Scientific Research at Princess Nourah bint Abdulrahman University through the Fast-track Research Funding Program.