Suicide and Depression Detection in Social Media Forums

Abstract

Suicide has been the first world issue for a long time. Depression is one of the major causes of suicide. There are various helpline centers and therapy sessions available for people who have a mental illness but are resistant to them since they have to speak openly about their problems to a person directly. So, people tend to open up on social media platforms often to hide their identities. These platforms are seeing a lot of texts related to suicides. The motive of our work is to use these texts from social media platforms and help identify the persons who need immediate help. The dataset used is collected from “SuicideWatch” and “depression” subreddits of the Reddit social media platform. The proposed method uses Apache Spark to implement different machine learning and deep learning algorithms like XGBoost, SVM, and artificial neural network (ANN) on a multi-node Spark cluster.