Framework for suicide detection from Arabic tweets using deep learning

Abstract

Major depressive disorder (MDD), is considered as a severe and widespread disease that causes suicide in many cases. This affects the thoughts, conduct, and quality of life of many people around the world. When treatment is not sought, suicide is regarded as the second most common cause of death. Due to people communicating their feelings and ideas on social media (Twitter) regarding a variety of topics in these tweets suicide can be predicted in advance. This study is one of many that advise tracking depression and other mental diseases using social media Arabic data. Arabic is a widely spoken language with difficult grammar; hence depression detection methods have not been widely used. Most of all previous studies were found in English for English language tweets. To deal with Arabic language tweets for the proposed research, tweets are collected in Arabic and then annotated by analytical experts in psychoanalysis. Moreover, different deep learning algorithms were used for training on this dataset. These were used to predict suicide cases in advance. The obtained result has an accuracy greater than 90%.